diff --git a/.gitignore b/.gitignore index 32963245..09b70da8 100644 --- a/.gitignore +++ b/.gitignore @@ -60,3 +60,6 @@ spaces/ # APA Summit 2027 phygital calls: the editors' PDFs, page copies and extracted text stay local submissions/phygital_*/cfp/source/ + +# Calls-for-papers raw pages for every submission arm (public repo; refetchable) +submissions/*/cfp/source/ diff --git a/submissions/CALENDAR.md b/submissions/CALENDAR.md index f0a01dd6..8a1bf389 100644 --- a/submissions/CALENDAR.md +++ b/submissions/CALENDAR.md @@ -15,14 +15,35 @@ place.** Update it when a date moves, and say what moved it. | 2026-09-10 | Full manuscript | OS/OT PDW Lima | yes | | 2026-10-01 | First draft for editor feedback | *Slacker* collection, Bloomsbury | optional | | 2026-10-05→07 | Attend the workshop, in person, Lima | OS/OT PDW Lima | yes — plus roundtable reviewing | +| 2026-10-27 | Accountability in the Triad | FAccT 2027 ([`facct2027_triad_accountability/`](facct2027_triad_accountability/)); abstract registration 27 Oct AoE; full 14-page paper 3 Nov AoE | **abstract registered 2026-10-26** (non-archival); full paper due 3 Nov AoE | | 2026-10-28→31 | Talk: deck, script and a live session ([`triadic_reduction/`](triadic_reduction/)) | ALGOCON, algorithmacy.org, Port of Spain | yes — slot day not yet fixed; added 2026-09-24. 20 minutes; 30 requested 2026-09-25, not confirmed | +| ~~2026-10-30~~ | Algorithmacy as an IS construct (8-page short paper) | ECIS 2027 ([`ecis2027_algorithmacy_is/`](ecis2027_algorithmacy_is/)); 30 Oct | **submitted 2026-10-26** (PCS #1520); decision 1 Feb 2027 | +| 2026-10-31 | Co-optation as Control (350–500-word abstract) | ILPC 2027 ([`ilpc2027_cooptation_control/`](ilpc2027_cooptation_control/)); **submitted 2026-10-26** (General stream); decisions early Dec | yes — abstract submitted | | 2026-11-01 | Final draft | *Slacker* collection, Bloomsbury | yes | +| 2026-11-01 | Governed Without a Seat (about 1,000-word abstract; MC09 or Network J) | SASE 2027 ([`sase2027_platform_cooptation/`](sase2027_platform_cooptation/)); 1 Nov (MC09); 16 Dec (Network J) | no — project opened 2026-10-26, nothing submitted | +| 2026-11-01 | an exact test for triadic coordination (proposal plus repository) | *Organization Science* AI-Enabled Frontiers ([`orgsci_ai_frontiers_methods/`](orgsci_ai_frontiers_methods/)); 1 Nov | no — project opened 2026-10-26, nothing submitted | | 2026-11-10 | 500-word abstract, APA Summit portal ([`phygital_amj_marketing/`](phygital_amj_marketing/)) | *Australasian Marketing Journal* phygital special issue | no — project opened 2026-10-23, not committed | +| 2026-11-15 | After Algorithmic Literacy (300–500-word abstract) | AMS 2027 (AlgorLit) ([`algorlit2027_literacy_to_algorithmacy/`](algorlit2027_literacy_to_algorithmacy/)); 15 Nov, 23:59 Madrid | no — project opened 2026-10-26, nothing submitted | +| 2026-11-15 | The Internal Third Party (full paper) | *JITCAR* The Post AI Enterprise ([`jitcar_post_ai_enterprise/`](jitcar_post_ai_enterprise/)); 15 Nov | no — project opened 2026-10-26, nothing submitted | +| 2026-12-15 | The Lived Arbiter (150–250-word abstract plus biography) | SPM2027 ([`spm2027_lived_arbiter/`](spm2027_lived_arbiter/)); 15 Dec | no — project opened 2026-10-26, nothing submitted | | 2026-12-31 | Submission-ready draft ([`algorithmacy_ot/`](algorithmacy_ot/)) | *Organization Theory*, regular theory article — algorithmacy | no — the author's own target, set 2026-10-07 | +| 2026-12-31 | Accountability Needs Three Places (600–800-word abstract; rolling) | CEPE 2027 ([`cepe2027_third_party_accountability/`](cepe2027_third_party_accountability/)); 31 Dec (rolling) | no — project opened 2026-10-26, nothing submitted | +| 2027-01-12 | the OT paper condensed for ST09_03 (full paper) | EURAM 2027 ([`euram2027_algorithmacy/`](euram2027_algorithmacy/)); 12 Jan, 14:00 Brussels | no — project opened 2026-10-26, nothing submitted | +| 2027-01-15 | Playing by Rules No One Wrote (alternate to the sub-theme 03 short paper) | EGOS 2027 sub-theme 64 ([`egos2027_st64_rules_of_play/`](egos2027_st64_rules_of_play/)); 15 Jan, 12:00 CET | no — project opened 2026-10-26, nothing submitted | | 2027-01-31 | Submission | *Organization Theory* Special Themed Section | yes | +| 2027-01-31 | Designing the Third Party | *ISJ* AI and the Future of Work ([`isj_future_of_work_design/`](isj_future_of_work_design/)); 31 Jan (indexed; verify) | no — project opened 2026-10-26, nothing submitted | +| 2027-01-31 | Selling Literacy for a Triadic Problem | *Information and Organization* Hype Studies ([`io_hype_ai_literacy/`](io_hype_ai_literacy/)); 31 Jan (aggregator; verify) | no — project opened 2026-10-26, nothing submitted | +| 2027-01-31 | *Information and Organization* Algorithmic Assemblages (stub; call apparently closed 15 May 2026) | *Information and Organization* Algorithmic Assemblages (stub; call apparently closed 15 May 2026) ([`io_algorithmic_assemblages/`](io_algorithmic_assemblages/)); unconfirmed | no — project opened 2026-10-26, nothing submitted | +| 2027-02-01 | Competence and Standing | *Organization Science* Agency in the Age of AI ([`orgsci_agency_triad/`](orgsci_agency_triad/)); 1 Feb | no — project opened 2026-10-26, nothing submitted | +| 2027-03-01 | When Your Agent Is Also Your Counterpart's | *MISQE* Delegation to AI Agents ([`misqe_delegation_triads/`](misqe_delegation_triads/)); 1 Mar | no — project opened 2026-10-26, nothing submitted | +| 2027-05-03 | Truth by Arbitration | *JMS* Organizing Truth ([`jms_organizing_truth_arbiter/`](jms_organizing_truth_arbiter/)); 3 May | no — project opened 2026-10-26, nothing submitted | | 2027-05-30 | Full manuscript ([`phygital_smq_value_creation/`](phygital_smq_value_creation/)) | *Services Marketing Quarterly* phygital special issue | no — not committed; eligibility without a Summit abstract unconfirmed | +| 2027-06-30 | Where Accountability Breaks in the Service Triad | *Service Industries Journal* ([`sij_service_triad_ethics/`](sij_service_triad_ethics/)); 30 Jun | no — project opened 2026-10-26, nothing submitted | | 2027-07-10 | Full manuscript ([`phygital_cjas_leadership/`](phygital_cjas_leadership/)) | *Canadian Journal of Administrative Sciences* phygital special issue | no — not committed; eligibility unconfirmed | | 2027-08-05 | Full manuscript ([`phygital_amj_marketing/`](phygital_amj_marketing/)) | *Australasian Marketing Journal* phygital special issue | no — not committed | +| 2027-08-27 | Algorithmacy at Work (needs data) | *Personnel Psychology* Generative Workplace ([`perspsych_measuring_algorithmacy/`](perspsych_measuring_algorithmacy/)); 1–27 Aug | no — project opened 2026-10-26, nothing submitted | + +**New crunch, 27 Oct – 1 Nov 2026:** five call deadlines in six days (FAccT registration, ECIS, ILPC, SASE MC09, *Organization Science* AI-Enabled Frontiers), on top of ALGOCON (28–31 Oct) and the *Slacker* final draft (1 Nov). None is committed; the author chooses which to pursue. **The crunch is 30 August to 10 September: three binding deadlines in twelve days.** The Lima place is confirmed and its manuscript already drafted, so the pressure sits on the IGI chapter and the diff --git a/submissions/CFP_AUDIT_2026-10-26.md b/submissions/CFP_AUDIT_2026-10-26.md new file mode 100644 index 00000000..4470dda5 --- /dev/null +++ b/submissions/CFP_AUDIT_2026-10-26.md @@ -0,0 +1,64 @@ +# Cross-arm audit of the 26 Oct 2026 CFP arms + +[DRAFTED by Claude, 26 Oct 2026. A condensed record of a read-only audit run on the Fable model, plus a citation audit (Sonnet) and a mechanical check (Haiku). Following the author's rule against AI-review → AI-rewrite loops, **the prose findings below were not applied**. They wait for the author's read-aloud. Only factual and rule fixes were made: +- the citation corrections listed in each arm's files; +- "AI drafts the appraisal" restated as a possibility (the source, Hinds & von Krogh, is conjecture); +- the FAccT abstract's "freelancer" example replaced; +- word counts and unresolvable paths corrected.] + +## Rulings needed once, for all arms +| # | Ruling | Arms affected | +|---|---|---| +| a | Which arm owns the "three-place forum" claim | FAccT, CEPE | +| b | Which arm owns "competence plus collective voice" (the OT paper already says exit is costly because ratings are lost) | ILPC, SASE | +| c | Which arm owns "competence plus standing" | *Organization Science* Agency, Exetasis | +| d | One vocabulary for the three facets. The OT paper says interpreting, specifying intent and tracking; the scale's codebook says counterpart inference, signal compression and rule-change tracking. | ECIS, *Personnel Psychology*, codebook | +| e | One home each for the moral crumple zone, the two failure modes and the Selznick reversal | CEPE, *Organization Science* Agency, EGOS 64, EURAM, ILPC, SASE | +| f | EGOS: sub-theme 03 (the author's Google Doc) or sub-theme 64. One short paper per main author. | EGOS 64 | +| g | Accept or reject the EURAM risk. The EURAM paper is the OT paper condensed, and EURAM bars work "published or accepted" before June 2027. | EURAM | +| h | How to cite the OT paper while it is under blind review | every spinoff | + +## Overlaps to separate +| Pair | Overlap | Proposed split | +|---|---|---| +| FAccT / CEPE | Near-identical conclusion (a forum both parties can call on) | FAccT owns two audit tests: joint contest of one verdict, and explanations that separate the counterpart's input from the rule's. CEPE owns Peirce's irreducibility and the crumple zone with two candidates. | +| ILPC / SASE | Same second half of the conclusion, sources and fused-score mechanism. The idea is also in the OT paper's text. | ILPC owns the relocation of the frontier of control and the counterpart as ally. SASE owns the four features ("not a fourth logic"), stratification and the masking effect. | +| *JITCAR* / *MISQE* | Both allocate the practices to roles; same "manager's summary" vignette | *MISQE* rests only on agent autonomy and drops the summary vignette. *JITCAR* owns the operator-plus-principal position. | +| *ISJ* / *SIJ* | *SIJ*'s guest-visible attribution and reciprocal reveal are interface features (*ISJ*'s territory) | *SIJ* keeps governance mechanisms: appeals, managerial adjustment, portability and oversight authority. | +| AlgorLit / *I&O* Hype | Same literacy-myth premise and closing caution | AlgorLit owns the definitional reading. *I&O* Hype owns the literacy myth and the remedy market. | +| SPM / EGOS 64 / CEPE / *ISJ* / OT | Simmel's mediator–arbitrator contrast and the "double address" recur | SPM claims only the lived reading. The others cite the double address and do not restate it. | +| EGOS 64 / OT | The two failure modes are the OT paper's | EGOS 64 builds on the convenors' question with anticipatory compliance as new evidence. | + +## Concurrency risks (ranked) +1. **EURAM:** the paper is the OT paper. A breach occurs if OT accepts it before June 2027. +2. **ECIS:** its concurrency clause covers similar versions, and accepted papers go into the proceedings. +3. **FAccT:** the archival choice is irreversible at registration (27 Oct) and must be non-archival. +4. **ILPC and SASE together** are one paper under two frames. +5. **EGOS:** sub-theme 03 against sub-theme 64. +6. **CEPE's journal route** would publish the forum argument first. +7. **MISQE:** safe only if it stays apart from *JITCAR*. + +## Prose findings (not applied; for the author's read-aloud) +- **Claim buried behind a question or setup:** *Organization Science* Agency, JMS, AlgorLit, FAccT (claim in sentence 6), CEPE (end of paragraph 2), EURAM (its opener echoes the OT abstract). +- **Epigram or antithesis closers:** SASE, EGOS 64, *JITCAR*, SPM, *SIJ*, *Organization Science* methods, *ISJ*, JMS, ILPC. +- **Hedge in prose:** CEPE ("Whatever the final status of his reduction thesis…"). +- **Self-narrated structure:** CEPE ("The argument has three steps"), *ISJ*, *Organization Science* methods, SASE, FAccT ("We close with…"). +- **Agentless passives:** JMS ("Its rules are withheld…"), *Personnel Psychology* ("were fixed before data collection"). +- **OT sentences reused in substance:** ILPC, SASE and EURAM each restate the OT paper's deskilling paragraph. + +## FLAGS gaps +- **ALGOCON (28–31 Oct) and the *Slacker* final draft (1 Nov)** collide with FAccT, ECIS, ILPC, SASE and *Organization Science* methods. Those FLAGS files do not say so. +- **AlgorLit:** if the key-dates calendar link (15 Oct) is right rather than the page text (15 Nov), the call has closed. +- **JMS:** an AI-use declaration is required. +- **Exetasis:** the P/P/C and outline drafted from the v4 abstract were withdrawn, because the arm moved to v5.2 (with Antonio Scala) on its own branch. A Coeckelbergh year mismatch (`coeckelbergh2022` in VERIFICATION.md against Coeckelbergh 2023 in the abstract) is still open there. +- **Burrell (2016):** at abstract depth, it carries a premise in FAccT, *ISJ*, SPM and EURAM. + +## Decisions by date +| By | Decision | +|---|---| +| 27 Oct AoE | FAccT: register (non-archival) or drop. | +| 30 Oct | ECIS: go or no-go given the concurrency clause. If go, choose the track. | +| 31 Oct | ILPC: go or no-go (the platform shows 8 Nov). | +| 1 Nov | SASE: MC09 now, or Network J by 16 Dec. *Organization Science* AI-Enabled Frontiers: go or no-go. | +| This week | Verify in a browser: the AlgorLit date, the *I&O* Hype PDW, and the blocked *ISJ*, *I&O*, *Personnel Psychology* and *Organization Science* pages. | +| Before 15 Nov | *JITCAR*: pursue or hold. | diff --git a/submissions/README.md b/submissions/README.md index 84b1fd10..49ef1398 100644 --- a/submissions/README.md +++ b/submissions/README.md @@ -27,5 +27,31 @@ This week's work is listed first in [`../NOW.md`](../NOW.md). | [`phygital_bh_transformation/`](phygital_bh_transformation/) | *Business Horizons* APA 2027 special issue: Phygital Business Transformation. Project opened; abstract and registration deadlines passed | not stated | | [`phi_program_review/`](phi_program_review/) | Synthesis review: five months of exact Φ on small coordination models, with the PyPhi pin and the upstream 2.0 control check | no deadline | +## Calls for papers opened 26 Oct 2026, by due date + +[DRAFTED by Claude, 26 Oct 2026.] Each folder holds a CFP digest, an abstract, a premise/premise/conclusion argument, an outline and a flags list, all marked as Claude's drafts for the author's ruling. Nothing has been submitted. The search behind them is summarised in each `cfp/README.md`. The *Exetasis* arm (20 Dec 2026) is not part of this batch: it moved to a v5.2 abstract with Antonio Scala's plan on its own branch (PR #851). Overlaps, concurrency risks and the rulings needed are in [`CFP_AUDIT_2026-10-26.md`](CFP_AUDIT_2026-10-26.md). + +| Due | Path | Piece | +| --- | --- | --- | +| abstract registration 27 Oct AoE; full 14-page paper 3 Nov AoE | [`facct2027_triad_accountability/`](facct2027_triad_accountability/) | FAccT 2027: Accountability in the Triad | +| ~~30 Oct~~ **submitted 26 Oct** | [`ecis2027_algorithmacy_is/`](ecis2027_algorithmacy_is/) | ECIS 2027: Algorithmacy as an IS construct (8-page short paper) | +| ~~31 Oct~~ **submitted 26 Oct** | [`ilpc2027_cooptation_control/`](ilpc2027_cooptation_control/) | ILPC 2027: Co-optation as Control (350–500-word abstract) | +| 1 Nov (MC09); 16 Dec (Network J) | [`sase2027_platform_cooptation/`](sase2027_platform_cooptation/) | SASE 2027: Governed Without a Seat (about 1,000-word abstract; MC09 or Network J) | +| 1 Nov | [`orgsci_ai_frontiers_methods/`](orgsci_ai_frontiers_methods/) | *Organization Science* AI-Enabled Frontiers: an exact test for triadic coordination (proposal plus repository) | +| 15 Nov, 23:59 Madrid | [`algorlit2027_literacy_to_algorithmacy/`](algorlit2027_literacy_to_algorithmacy/) | AMS 2027 (AlgorLit): After Algorithmic Literacy (300–500-word abstract) | +| 15 Nov | [`jitcar_post_ai_enterprise/`](jitcar_post_ai_enterprise/) | *JITCAR* The Post AI Enterprise: The Internal Third Party (full paper) | +| 15 Dec | [`spm2027_lived_arbiter/`](spm2027_lived_arbiter/) | SPM2027: The Lived Arbiter (150–250-word abstract plus biography) | +| 31 Dec (rolling) | [`cepe2027_third_party_accountability/`](cepe2027_third_party_accountability/) | CEPE 2027: Accountability Needs Three Places (600–800-word abstract; rolling) | +| 12 Jan, 14:00 Brussels | [`euram2027_algorithmacy/`](euram2027_algorithmacy/) | EURAM 2027: the OT paper condensed for ST09_03 (full paper) | +| 15 Jan, 12:00 CET | [`egos2027_st64_rules_of_play/`](egos2027_st64_rules_of_play/) | EGOS 2027 sub-theme 64: Playing by Rules No One Wrote (alternate to the sub-theme 03 short paper) | +| 31 Jan (indexed; verify) | [`isj_future_of_work_design/`](isj_future_of_work_design/) | *ISJ* AI and the Future of Work: Designing the Third Party | +| 31 Jan (aggregator; verify) | [`io_hype_ai_literacy/`](io_hype_ai_literacy/) | *Information and Organization* Hype Studies: Selling Literacy for a Triadic Problem | +| unconfirmed | [`io_algorithmic_assemblages/`](io_algorithmic_assemblages/) | *Information and Organization* Algorithmic Assemblages (stub; call apparently closed 15 May 2026) | +| 1 Feb | [`orgsci_agency_triad/`](orgsci_agency_triad/) | *Organization Science* Agency in the Age of AI: Competence and Standing | +| 1 Mar | [`misqe_delegation_triads/`](misqe_delegation_triads/) | *MISQE* Delegation to AI Agents: When Your Agent Is Also Your Counterpart's | +| 3 May | [`jms_organizing_truth_arbiter/`](jms_organizing_truth_arbiter/) | *JMS* Organizing Truth: Truth by Arbitration | +| 30 Jun | [`sij_service_triad_ethics/`](sij_service_triad_ethics/) | *Service Industries Journal*: Where Accountability Breaks in the Service Triad | +| 1–27 Aug | [`perspsych_measuring_algorithmacy/`](perspsych_measuring_algorithmacy/) | *Personnel Psychology* Generative Workplace: Algorithmacy at Work (needs data) | + These arms borrow lab constructs. They do not re-run Φ. Exact-Φ results remain evidence about Boolean models. diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/ABSTRACT.md b/submissions/algorlit2027_literacy_to_algorithmacy/ABSTRACT.md new file mode 100644 index 00000000..9bd52fca --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/ABSTRACT.md @@ -0,0 +1,27 @@ +# Abstract: AMS 2027 (AlgorLit), individual paper + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. 300–500 words, covering objectives, methodology, results and contribution. Due 15 Nov 2026, 23:59 Madrid.] + +**Title:** After Algorithmic Literacy: Coordinating With Another Person Through an Algorithm + +## Abstract +**Objectives.** Algorithmic literacy has become the field's name for what people need in algorithmic environments. This paper asks whether that name fits a common situation it was not built for. In that situation, two people coordinate through a system that evaluates both of them: a host and a guest under a booking platform's ranking, or a seller and a buyer under a marketplace's ratings. It also asks what the history of literacy suggests about promising competence as a remedy. + +**Methodology.** The paper is a conceptual analysis in two parts. First, it reads the definitions collected in the field's two integrative reviews, one of 50 articles and one of 169 studies, together with the leading conceptual and measurement proposals. For each definition it records who the parties are and what the person is said to know or do. Second, it compares this pattern with historical scholarship on print-literacy campaigns and with a recent critique of AI-literacy initiatives. + +**Results.** In every definition examined, the relation is between one user and one system. The dimensions describe a person's awareness of algorithms, knowledge of how they work, critical evaluation of them, and ways of coping with or influencing them. None describes a second person on the other side of the system. Yet much algorithmic power runs between people. A host's ranking moves with guests' reviews, and a guest's choice moves with the ranking. The historical comparison shows a second pattern. Literacy campaigns promised that individual competence would bring advancement, and literacy was often more valuable to the society's goals than to most individuals. AI-literacy initiatives use literacy as a metaphor and repeat this promise. + +**Contribution.** The paper proposes a structural distinction. Oracy names competence in direct exchange, and literacy names competence with a text that a human still interprets and judges. Algorithmacy names competence in coordinating with another person through a system that evaluates both. It has three practices: interpreting an ambiguous result to tell the counterpart's part from the system's, specifying intent so the counterpart can read it through the system, and tracking rule changes the operator does not announce. For algorithmic-literacy research, the contribution is a new object of study, the counterpart. For literacy education, it is a caution: knowing how the system works does not give a person standing in front of it, so programmes should not promise that it will. + +**Keywords:** algorithmic literacy; algorithmacy; platforms; literacy myth; coordination + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Review of 50 articles; journal pagination | `dxd_foundation/literature/library/oeldorfhirsch2023what.md` (p. 683) | full text | +| Review of 169 studies | `lima_pdw/literature/cards/gagrcin2026.md` | full text | +| Four dimensions (awareness and knowledge, evaluation, coping, creation) | `lima_pdw/literature/cards/dogruel2021.md` | extended | +| Algorithmic literacy definition | `lima_pdw/literature/cards/devito2021.md` | extended | +| Literacy more valuable to society's goals than to most individuals | `berthon_ethics_course/literature/cards/graff1991literacy.md` (p. 321) | partial text | +| AI literacy used "uncritically as a metaphor" | `berthon_ethics_course/literature/cards/kamperman2026ai.md` (p. 35) | full text | +| Host–guest ranking dynamics | `dxd_foundation/literature/library/jhaver2018algorithmic.md` | full text | diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/ARGUMENT.md b/submissions/algorlit2027_literacy_to_algorithmacy/ARGUMENT.md new file mode 100644 index 00000000..4b25fb8e --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/ARGUMENT.md @@ -0,0 +1,18 @@ +# Argument: from algorithmic literacy to algorithmacy + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Algorithmic literacy is defined for a person facing a system. Its dimensions (awareness, knowledge, critical evaluation, coping, creation) describe what one user knows and does about an algorithm. The field's two reviews find no cohesive construct and no human counterpart in the relation studied. +- *Sources:* + - `lima_pdw/literature/cards/dogruel2021.md` (four dimensions) + - `lima_pdw/literature/cards/devito2021.md` (definition) + - `dxd_foundation/literature/library/oeldorfhirsch2023what.md` (journal pagination; p. 693) + - `lima_pdw/literature/cards/gagrcin2026.md` (169 studies; no cohesive framework) + +**P2.** The history of literacy shows that a competence named after the medium a person reads tells us little about the social relation the medium sets up. Literacy campaigns promised individual advancement that the social order did not deliver, and AI-literacy initiatives repeat that promise. +- *Sources:* + - `berthon_ethics_course/literature/cards/graff1991literacy.md` (pp. 2, 321) + - `berthon_ethics_course/literature/cards/kamperman2026ai.md` (pp. 35–36) + - `berthon_ethics_course/literature/cards/ong1982orality.md` (p. 133) + +**C.** Where an algorithm stands between two people and evaluates both, the competence that matters is not literacy about the algorithm. It is algorithmacy: the competence to coordinate with the other person through it. Algorithmic-literacy research should add the counterpart to its object, and literacy programmes should stop promising that knowledge of the system will deliver standing. diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/FLAGS.md b/submissions/algorlit2027_literacy_to_algorithmacy/FLAGS.md new file mode 100644 index 00000000..641056bf --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: AlgorLit 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The template asks for "Results".** For a conceptual paper, the results are the reading of the definitions. That reading is the author's: "none describes a second person" is a claim about the definitions examined, and it must be checked against each one before submission. +2. **Oeldorf-Hirsch & Neubaum** focus on social media, so the absence of a counterpart there is partly a matter of scope (card caution). +3. **Gagrčin et al.'s 169 studies** include 5 dissertations and 78 found by citation chasing, so do not call them "169 peer-reviewed". +4. **Graff's pages** are from the 1991 reissue. Graff also calls the myth "partly true". +5. **Kamperman & Graff's own remedy** is person-side (curriculum and disclosure, p. 37). Do not present them as allies of the structural argument. +6. **Wilkinson (1965), Olson (1994) and Goody & Watt (1963)** are not read. They are not cited in the abstract. +7. **Double-blind review:** the abstract does not name the author's prior work. +8. **Confirm the date.** The page's calendar link says 15 October. diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/OUTLINE.md b/submissions/algorlit2027_literacy_to_algorithmacy/OUTLINE.md new file mode 100644 index 00000000..f77a1001 --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/OUTLINE.md @@ -0,0 +1,13 @@ +# Outline: full paper after acceptance (4,000–6,000 words, APA 7) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 600 | Gives the host–guest case and asks whether literacy fits it. | `jhaver2018algorithmic` | +| 2 | What algorithmic literacy names | 1,200 | Reads the definitions for the parties and the person's acts. | `dogruel2021`, `devito2021`, `oeldorfhirsch2023what`, `gagrcin2026`, `cotterreisdorf2020` (abstract), `swart2021` (abstract) | +| 3 | What the literacy campaigns promised | 1,000 | Covers the literacy myth and AI-literacy marketing. | `graff1991literacy`, `kamperman2026ai`, `goody1987interface` (partial) | +| 4 | Oracy, literacy, algorithmacy | 1,300 | Gives the structural criterion and the three practices. | `ong1982orality` (p. 133), `wilkinson1965` (metadata), `simmel1950sociology` | +| 5 | Implications | 800 | For research (the counterpart as object) and for education (no promise of standing). | `oeldorfhirsch2023what` (p. 693) | +| 6 | Conclusion | 300 | — | — | +| | **Total** | **5,200** | | | diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/README.md b/submissions/algorlit2027_literacy_to_algorithmacy/README.md new file mode 100644 index 00000000..397f52d3 --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/README.md @@ -0,0 +1,18 @@ +# AMS 2027 (AlgorLit): After Algorithmic Literacy + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **The abstract (300–500 words) is due 15 Nov 2026, 23:59 Madrid.** Nothing has been submitted. + +**The call:** AMS International Conference 2027, Madrid, 24–25 May 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date | +|---|---| +| Abstract | **15 Nov 2026, 23:59 Madrid** | +| Notification | 15 Jan 2027 | +| Confirmation | before 1 Mar 2027 | +| Full paper (4,000–6,000 words) | after acceptance; date not stated | + +**Angle:** algorithmic literacy is defined for a user facing a system, while algorithmacy covers coordinating with a counterpart through it. The literacy myth is a caution. + +**Relation to other arms:** draws on [`../berthon_ethics_course/`](../berthon_ethics_course/) (literacy history). It is distinct from [`../io_hype_ai_literacy/`](../io_hype_ai_literacy/), which studies the market for AI-literacy remedies. diff --git a/submissions/algorlit2027_literacy_to_algorithmacy/cfp/README.md b/submissions/algorlit2027_literacy_to_algorithmacy/cfp/README.md new file mode 100644 index 00000000..c7b6f2e5 --- /dev/null +++ b/submissions/algorlit2027_literacy_to_algorithmacy/cfp/README.md @@ -0,0 +1,22 @@ +# CFP brief: AMS 2027, International Conference on Algorithmic Literacy (AlgorLit) + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** algorlit.com/en/congress/ams-2027/ (the about, participation-types, submission-guidelines, general-guidelines, key-dates and thematic-areas pages). The call PDF was unreadable. + +**Event:** AMS International Conference 2027 ("Algorithms, Media & Society"). Universidad CEU San Pablo, Madrid, 24–25 May 2027. In person, in Spanish and English. Organised by the AlgorLit Project. + +## Dates (verbatim) +- Submission: "15 November 2026 · 23:59 Madrid time". "Abstracts submitted after the deadline will not be considered." +- Notification: "15 January 2027". Confirmation: "Before 1 March 2027". +- The key-dates page's calendar link encodes 15 October, but the page text says 15 November. + +## Format (verbatim) +- Individual paper abstract: "Between 300 and 500 words". Keywords: "Between 3 and 5". The abstract "MUST COVER": "Objectives", "Methodology", "Results", "Contribution to the field". +- Full paper after acceptance: "4,000 – 6,000 words", .docx, Times New Roman 12, 1.5 spacing, APA 7. +- "Double-blind peer review". Limit: "1 as lead author + 1 as co-author". Submit only through the online form. + +## Lines nearest this paper (verbatim) +- "algorithmic literacy emerges as a key competence for understanding, questioning and exercising critical citizenship in digital environments" +- Thematic area 01: "Critical algorithmic literacy"; "Education and algorithmic awareness". +- Thematic area 03: "Algorithmic governance"; "Transparency and accountability"; "Platform power and responsibility". diff --git a/submissions/cepe2027_third_party_accountability/ABSTRACT.md b/submissions/cepe2027_third_party_accountability/ABSTRACT.md new file mode 100644 index 00000000..cd5d321a --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/ABSTRACT.md @@ -0,0 +1,41 @@ +# Abstract: CEPE 2027 (600–800 words, APA) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Send to cepe2027@ihrc.gr by 31 Dec 2026; earlier is better, since review is rolling.] + +**Title:** Accountability Needs Three Places: AI Systems That Judge Two People + +## Abstract +Algorithmic accountability has inherited a two-place form. Bovens (2007) defines accountability as a relation between an actor and a forum: the actor must explain and justify its conduct, and the forum can question it, pass judgement and impose consequences. Proposals for accountable AI keep this form when they turn to the people a system affects. Contestability frameworks give a decision subject the chance to challenge a decision before a human controller (Alfrink et al., 2023). Stakeholders, asked what contestability requires, mostly describe an appeal after the fact (Lyons et al., 2021). In each case one person faces one system, or the people who answer for it. + +Many AI systems do not stand in that relation. A booking platform ranks a host partly on guests' reviews, and a marketplace demotes a seller on buyers' ratings. An AI appraisal tool can draft what a manager tells an employee. In each case the system evaluates the conduct of two people who depend on each other, and its verdict binds both. This paper argues that such systems stand in a three-place relation, and that two-place accountability, applied to them, misplaces responsibility. + +The argument has three steps. The first is logical. Peirce held that a genuinely triadic relation cannot be analysed into dyadic ones: "a triad cannot be analyzed into dyads" (Peirce, c. 1890, CP 1.363). Whatever the final status of his reduction thesis, which later work shows to depend on how definability is construed, it gives a precise test. A verdict about two people's joint conduct is triadic if it cannot be rebuilt from two separate verdicts, one about each person. A ranking that combines a guest's review with the platform's weighting of it fails that test. So does a seller's standing that depends on a buyer's rating and on the operator's rule. + +The second step is sociological. Simmel's analysis of the third shows what changes when a third decides. The arbitrator ends the parties' control over their own conflict. A third can also benefit from the parties' division, and can produce that division in order to dominate (Simmel, 1950, pp. 154, 162). Platforms occupy these positions. In Stark and Vanden Broeck's (2024, p. 14) terms, they leverage the agency of participants rather than delegating to them, and accountability becomes "distributed, deflected, and denied". + +The third step is normative. Two-place mechanisms hear each party alone. A host may contest a ranking, and a guest may complain about a host, but no forum hears the joint conduct that the verdict judged. Responsibility then settles on whichever party is nearest the outcome. Elish (2019) calls this the moral crumple zone, in which the nearest human absorbs the blame for a system over which they had limited control. In a triad the crumple zone has two candidates, and the operator can choose between them. Better explanations do not remedy this, because disclosure is not the same as traceability (Kroll, 2021), and an explanation given to one party still leaves the other unheard. Nor does a right of appeal, which lets each party reopen its own verdict but not the joint one. The host who wins an appeal against a ranking leaves the guest's review, and the platform's weighting of it, exactly where they were. + +The paper concludes that accountability for AI systems that judge two people needs a three-place form. It needs a forum before which both parties and the operator answer together, with a record that both parties can read. It also needs a standard of justification that addresses the joint conduct, not each party's conduct separately. The paper sets out the minimal conditions for such a forum and asks which existing institutions come close: two-sided review disputes, mediation services in online dispute resolution, and labour tribunals. + +## References +- Alfrink, K., Keller, I., Kortuem, G., & Doorn, N. (2023). Contestable AI by design: Towards a framework. *Minds and Machines, 33*(4), 613–639. https://doi.org/10.1007/s11023-022-09611-z +- Bovens, M. (2007). Analysing and assessing accountability: A conceptual framework. *European Law Journal, 13*(4), 447–468. https://doi.org/10.1111/j.1468-0386.2007.00378.x +- Elish, M. C. (2019). Moral crumple zones: Cautionary tales in human-robot interaction. *Engaging Science, Technology, and Society, 5*, 40–60. https://doi.org/10.17351/ests2019.260 +- Kroll, J. A. (2021). Outlining traceability: A principle for operationalizing accountability in computing systems. In *Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency* (pp. 758–771). https://doi.org/10.1145/3442188.3445937 +- Lyons, H., Velloso, E., & Miller, T. (2021). Conceptualising contestability: Perspectives on contesting algorithmic decisions. *Proceedings of the ACM on Human-Computer Interaction, 5*(CSCW1), Article 106. https://doi.org/10.1145/3449180 +- Peirce, C. S. (c. 1890). A guess at the riddle. In C. Hartshorne & P. Weiss (Eds.), *Collected papers of Charles Sanders Peirce* (Vol. 1, CP 1.354–416). Harvard University Press. +- Simmel, G. (1950). *The sociology of Georg Simmel* (K. H. Wolff, Ed. & Trans.). Free Press. +- Stark, D., & Vanden Broeck, P. (2024). Principles of algorithmic management. *Organization Theory, 5*(2), 1–24. https://doi.org/10.1177/26317877241257213 + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Actor–forum definition | `lima_pdw/literature/cards/bovens2007.md` | full text (no page on card) | +| Dialectical exchange between decision subject and controller | `dxd_foundation/literature/library/alfrink2023contestable.md` (p. 615) | full text | +| Contestability as post-hoc appeal | `dxd_foundation/literature/library/lyons2021conceptualising.md` (p. 15, arXiv) | full text | +| "a triad cannot be analyzed into dyads" | `triadic_reduction/library/cards/peirce1890guess.md` (CP 1.363) | page images | +| The thesis depends on definability | `triadic_reduction/library/cards/herzberger1981theorem.md` | snippets only | +| The third who benefits and who divides | `dxd_foundation/literature/library/simmel1950sociology.md` (pp. 154, 162) | full text | +| "distributed, deflected, and denied"; leverages rather than delegates | `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) | full text | +| Moral crumple zone | `dxd_foundation/literature/library/elish2019moral.md` (pp. 40, 49) | full text | +| Disclosure is not traceability | `dxd_foundation/literature/library/kroll2021outlining.md` (p. 4, arXiv) | full text | diff --git a/submissions/cepe2027_third_party_accountability/ARGUMENT.md b/submissions/cepe2027_third_party_accountability/ARGUMENT.md new file mode 100644 index 00000000..ec4f00bf --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/ARGUMENT.md @@ -0,0 +1,20 @@ +# Argument: accountability needs three places + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Accountability, as a mechanism, is a two-place relation: an actor and a forum. The actor explains and justifies its conduct, and the forum questions it, judges it and imposes consequences. Algorithmic-accountability proposals keep this form. A decision subject contests the system's decision before a human controller or an appeal process. +- *Sources:* + - `lima_pdw/literature/cards/bovens2007.md` + - `lima_pdw/literature/cards/bovens2010.md` (virtue versus mechanism) + - `dxd_foundation/literature/library/alfrink2023contestable.md` (p. 615) + - `dxd_foundation/literature/library/lyons2021conceptualising.md` (p. 15, arXiv pagination) + +**P2.** An AI system that evaluates two interdependent people and binds both stands in a genuinely three-place relation, because its verdict is about their joint conduct. On Peirce's reduction thesis, such a relation cannot be built from two-place relations without loss. Simmel's sociology shows what is lost: a third who decides can also profit from the parties' division and keep them apart. +- *Sources:* + - `triadic_reduction/library/cards/peirce1890guess.md` (CP 1.363) + - `triadic_reduction/library/cards/peirce1870description.md` (p. 374) + - `dxd_foundation/literature/library/simmel1950sociology.md` (pp. 154, 162) + - `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) + +**C.** Two-place accountability, applied to a three-place relation, misplaces responsibility. Each party contests alone, the joint conduct is never heard as joint, and responsibility falls on whichever party is nearest, the moral crumple zone. Accountability for AI-mediated systems therefore needs a three-place form: a forum before which both parties and the system's operator answer together. +- *Source for the crumple zone:* `dxd_foundation/literature/library/elish2019moral.md` (pp. 40, 49). diff --git a/submissions/cepe2027_third_party_accountability/FLAGS.md b/submissions/cepe2027_third_party_accountability/FLAGS.md new file mode 100644 index 00000000..185f1375 --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: CEPE 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **Bovens (2007) is quoted in paraphrase only.** The card gives the definition without a page. Find the page in the PDF before quoting. +2. **Peirce's reduction thesis is contested.** Herzberger's card (snippets only) reports that it holds only under stated conditions. The abstract says so in a clause. Burch (1991) is metadata only. +3. **The CP 1.363 date:** the Peirce Edition Project dates "A guess at the riddle" 1887–88. The abstract says "c. 1890" with CP. +4. **Simmel's arbitrator is chosen** by the parties. The abstract does not say platforms are chosen. +5. **Alfrink and Lyons pages:** Alfrink's are journal pages; Lyons's and Kroll's are arXiv pages. Convert them before print. +6. **The "near-existing institutions" sentence** (two-sided review disputes, online dispute resolution, labour tribunals) is a list for the paper to examine. No card supports any of them yet. The Katsh & Rifkin "fourth party" cards exist but need their wording re-checked. +7. **Double-blind:** the abstract does not mention the author's other work. +8. **Overlap with `../facct2027_triad_accountability/`:** both argue three-place accountability. FAccT is computational and audit-facing; CEPE carries the logical and philosophical argument (Peirce, Simmel). If both are accepted, the second presented must cite the first. Decide which to pursue, or keep both distinct as above. diff --git a/submissions/cepe2027_third_party_accountability/OUTLINE.md b/submissions/cepe2027_third_party_accountability/OUTLINE.md new file mode 100644 index 00000000..387bc67e --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/OUTLINE.md @@ -0,0 +1,13 @@ +# Outline: CEPE full paper (at most 5,000 words, excluding bibliography) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Accountability's two places | 900 | Covers Bovens's actor–forum relation and its algorithmic uptake: contestability, appeal and explanation. | `bovens2007`, `bovens2010`, `alfrink2023contestable`, `lyons2021conceptualising`, `edwardsveale2017` | +| 2 | Systems that judge two people | 700 | Host and guest, seller and buyer, manager and employee. | `stark2020algorithmic` (p. 53), `jhaver2018algorithmic` | +| 3 | The logical step: Peirce | 900 | The reduction thesis, its contested status, and the test it gives. | `peirce1890guess`, `peirce1870description`, `peirce1897logic`, `herzberger1981theorem` (snippets) | +| 4 | The sociological step: Simmel | 800 | Arbitrator, *tertius gaudens* and *divide et impera*. | `simmel1950sociology`, `simmel1902number2` | +| 5 | The normative step | 1,000 | Misplaced responsibility, a crumple zone with two candidates, and why explanation does not fix it. | `elish2019moral`, `kroll2021outlining`, `starkvandenbroeck2024` | +| 6 | A three-place forum | 600 | Minimal conditions and near-existing institutions. | `katsh2001` (online dispute resolution "fourth party"; re-check wording first) | +| | **Total** | **4,900** | | | diff --git a/submissions/cepe2027_third_party_accountability/README.md b/submissions/cepe2027_third_party_accountability/README.md new file mode 100644 index 00000000..093c8f0d --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/README.md @@ -0,0 +1,18 @@ +# CEPE 2027: Accountability Needs Three Places + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages.] + +**Status:** project opened on 26 Oct 2026, with an abstract (600–800 words), argument and outline drafted. **Due 31 Dec 2026, by email to cepe2027@ihrc.gr; review is rolling and has started.** Nothing has been sent. + +**The call:** CEPE 2027, Athens, 26–28 May 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date | +|---|---| +| Abstract (600–800 words) | **31 Dec 2026 (rolling)** | +| Notification | 31 Jan 2027 | +| Full paper (≤5,000 words; only for the journal route) | 15 Mar 2027 | +| Registration | 1 Apr 2027 | + +**Angle:** accountability for AI systems that judge two people needs a three-place form (Peirce, Simmel, Bovens). + +**Relation to other arms:** shares its thesis with [`../facct2027_triad_accountability/`](../facct2027_triad_accountability/); see FLAGS 8. Draws on [`../triadic_reduction/`](../triadic_reduction/). diff --git a/submissions/cepe2027_third_party_accountability/cfp/README.md b/submissions/cepe2027_third_party_accountability/cfp/README.md new file mode 100644 index 00000000..fd980c12 --- /dev/null +++ b/submissions/cepe2027_third_party_accountability/cfp/README.md @@ -0,0 +1,29 @@ +# CFP brief: CEPE 2027 (Computer Ethics: Philosophical Enquiry) + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://events.ihrc.gr/cepe2027/call +- https://events.ihrc.gr/cepe2027/ +- https://philevents.org/event/show/154713 + +**Event:** CEPE 2027, "Ethics and Artificial Intelligence". UNIC Athens and Ionian University, Athens, 26–28 May 2027. Organised under INSEIT. Contact: cepe2027@ihrc.gr. + +## Dates (verbatim) +- "Abstracts / Panels / Roundtables / Workshops Submission Deadline: December 31, 2026" +- "Absracts of papers are reviewed on a rolling basis. The review has started." (sic) +- "Notification of Acceptance for abstracts: January 31, 2027" +- "Full Manuscript Due: March 15, 2027". "Submission of full papers is mandatory only for papers to be submitted to the international journals connected to the conference." +- PhilEvents: "Registration deadline: April 1, 2027, 9:00am EET". + +## Format (verbatim) +- "All abstracts should be written in English and prepared according to the APA guidelines and should not be less than 600 and not exceed 800 words, excluding bibliography. Papers should be no more than 5.000 words excluding bibliography." +- "Submissions should be sent at cepe2027@ihrc.gr" +- "All submissions will undergo double-blind peer review." + +## Lines nearest this paper (verbatim) +- "Accountability and AI-Mediated systems" +- "Computer ethics, power and responsibility" +- "Both theoretical and empirical contributions are encouraged." + +**Correction to the 26 Oct search:** the deadline is 31 Dec 2026, not "rolling only". Rolling review means earlier is better. diff --git a/submissions/ecis2027_algorithmacy_is/ABSTRACT.md b/submissions/ecis2027_algorithmacy_is/ABSTRACT.md new file mode 100644 index 00000000..ca142d9e --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/ABSTRACT.md @@ -0,0 +1,22 @@ +# Abstract: ECIS 2027 short paper (Track 4 or 17) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. ECIS states no abstract limit; this one is 142 words. The paper is due 30 Oct 2026.] + +**Title:** Algorithmacy: A Construct for Coordinating Through an Evaluating System + +## Abstract +IS research has several constructs for how people work with algorithms. These are AI literacy, algorithm literacy, algorithmic competency and algorithm sensemaking, and each places the system in one position. It is a partner to one user, a channel between two users, or a manager over one worker. Many platforms hold a fourth position: they stand between two interdependent people, evaluate both, and bind both to their verdicts. This short paper defines algorithmacy as the competence to coordinate with another person through such a system. It has three facets. The first is interpreting an ambiguous result, the second is specifying intent so the counterpart can read it through the system, and the third is tracking rule changes that go unannounced. I set out a research model that separates algorithmacy from its rivals, and state the scale development and validation work that remains. + +**Keywords:** algorithmacy; algorithmic management; AI literacy; coordination; construct development + +**Word count:** 142 (body, by `wc -w`). + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| AI literacy competencies | `lima_pdw/literature/cards/long2020.md` | full text | +| Algorithm literacy dimensions | `lima_pdw/literature/cards/dogruel2022.md`; `lima_pdw/literature/cards/gagrcin2026.md` | abstract; full text | +| Algorithmic competency | `lima_pdw/literature/cards/jarrahisutherland2019.md`; `lima_pdw/literature/cards/zhou2025apjhr.md` | abstract; full text | +| Algorithm sensemaking | `lima_pdw/literature/cards/mohlmann2023.md` | abstract | +| System as channel (AI-mediated communication) | `dxd_foundation/literature/library/hancock2020ai.md` | full text | +| Three-sided platform | `dxd_foundation/literature/library/stark2020algorithmic.md` (p. 53) | full text | diff --git a/submissions/ecis2027_algorithmacy_is/ARGUMENT.md b/submissions/ecis2027_algorithmacy_is/ARGUMENT.md new file mode 100644 index 00000000..d7243f04 --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/ARGUMENT.md @@ -0,0 +1,20 @@ +# Argument: Algorithmacy as an IS construct + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling. Sources are card paths in this repo.] + +**P1.** Each IS-adjacent construct for working with algorithms fixes one structural position for the system: a partner to one user, a channel between two users, or a manager over one worker. In none of them is the system an evaluator of two interdependent people. +- *Warrant:* the constructs' own definitions and items place the system that way. +- *Sources:* + - AI literacy: `lima_pdw/literature/cards/long2020.md` (full text) + - algorithm literacy: `lima_pdw/literature/cards/dogruel2022.md` (abstract) and `lima_pdw/literature/cards/gagrcin2026.md` (full text) + - algorithmic competency: `lima_pdw/literature/cards/jarrahisutherland2019.md` (abstract) and `lima_pdw/literature/cards/zhou2025apjhr.md` (full text) + - algorithm sensemaking: `lima_pdw/literature/cards/mohlmann2023.md` (abstract) + - AI-mediated communication: `dxd_foundation/literature/library/hancock2020ai.md` (full text) + +**P2.** On platforms and in AI-mediated work, the system often holds the third position. It evaluates both parties and binds both to its verdict, so each party's outcome depends on the counterpart and on the rule at once. +- *Warrant:* published studies of marketplaces and platforms describe a three-sided structure (Stark & Pais) in which one side evaluates the other through the operator (Curchod et al.). +- *Sources:* `dxd_foundation/literature/library/stark2020algorithmic.md` (p. 53); `dxd_foundation/literature/library/curchod2020working.md`; `lima_pdw/literature/cards/mohlmann2021.md` (matching and control). + +**C.** IS needs a construct defined by that third position. Algorithmacy, the competence to coordinate with another person through an evaluating intermediary, predicts outcomes that the existing constructs cannot. In particular, it predicts discriminant effects where the counterpart's behaviour and the rule's behaviour must be told apart. +- *Warrant for the step:* a construct is warranted when its defining structure differs from its rivals' and that difference implies testable relations they cannot produce. +- *Testable form:* the research model in OUTLINE §5. It has three facets: interpreting results, specifying intent and tracking rule change. It sets out discriminant relations with AI literacy and algorithmic competency, and its boundary conditions are task interdependence and perceived system authority. diff --git a/submissions/ecis2027_algorithmacy_is/FLAGS.md b/submissions/ecis2027_algorithmacy_is/FLAGS.md new file mode 100644 index 00000000..5a16e0e3 --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/FLAGS.md @@ -0,0 +1,20 @@ +# Flags: ECIS 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +## Deadline and format +1. **The full 8-page short paper is due 30 Oct 2026.** No timezone is stated. Four days remain. +2. **Concurrent consideration is prohibited:** "must not be concurrently under consideration for publication or presentation elsewhere." The OT paper's §2 already criticises the rival constructs. This paper must contribute the research model and discriminant logic (§§4–5), not restate the OT critique. If the author judges the overlap too close, ECIS should not go ahead while the OT paper is under review. +3. **The paper appears in the proceedings.** A later journal version of the same construct work must disclose this. +4. **Author limits:** one track only; at most three submissions per author. + +## Evidence +5. **Long & Magerko (2020), competency 10,** names humans who program and tune AI. Do not write "no competency names another person". Write that none names a counterpart the user coordinates with. +6. **Jarrahi & Sutherland 2019, Dogruel et al. 2022 and Möhlmann et al. 2023** are abstract depth only. +7. **Zhou trap:** never write "Zhou et al. 2025" bare. The scale is Zhou, Lei, Liu, Huang & Hou (*APJHR*). Do not invent page numbers, since Crossref gives none. +8. **Cite Dogruel et al. as 2022,** not 2021. The 2021 entry is a different chapter. +9. **No card exists for** Baird & Maruping 2021, Benbya, or Baskerville et al. 2020. They are not cited until read. +10. **The survey instrument** (`org_frontier/survey/cohort_algorithmacy/`) has never been fielded. Its pre-registration file has not been registered with an external registry, so call it "claims fixed before computation" if mentioned. + +## Fit +11. **Example rule:** no driver or freelancer examples lead the paper. Möhlmann's and Rahman's settings are cited for theory only. diff --git a/submissions/ecis2027_algorithmacy_is/OUTLINE.md b/submissions/ecis2027_algorithmacy_is/OUTLINE.md new file mode 100644 index 00000000..a7621ac7 --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/OUTLINE.md @@ -0,0 +1,16 @@ +# Outline: Algorithmacy as an IS construct (ECIS 2027 short paper, 8 pages all-inclusive) + +[DRAFTED by Claude, 26 Oct 2026. Eight single-spaced pages in the ECIS template hold about 4,200 words including references. Budget: about 3,300 words of text plus about 900 words of references.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 450 | Gives one case, a marketplace seller demoted after a buyer's review, and asks what competence the seller needs. Answers: competence defined by the system's position, not by a list of knowledge items. | `curchod2020working`, `stark2020algorithmic` | +| 2 | Four positions for the system | 800 | Partner (AI literacy), channel (AI-mediated communication), manager (algorithmic competency, algorithm sensemaking) and evaluator between two (the triad). One table sets construct, position, who is counted, and what each leaves out. | `long2020` (cite competencies by number; competency 10 names designers, see FLAGS), `dogruel2022`, `gagrcin2026`, `hancock2020ai`, `jarrahisutherland2019`, `zhou2025apjhr`, `mohlmann2023` | +| 3 | Algorithmacy defined | 600 | The definition, then three facets. Each facet answers one thing the evaluator removes: who caused a result, what a message meant, and whether the rule changed. | `stark2020algorithmic`, `rahman2021` (theory only), `devito2021` | +| 4 | Discriminant logic | 500 | Algorithmacy should diverge from AI literacy where knowledge of ML is high but the counterpart is ignored. It should diverge from algorithmic competency where the worker games the system but loses the counterpart. | `zhou2025apjhr` (items are worker-only), `bucherschouwaldkirch2021` (anticipatory compliance) | +| 5 | Research model | 550 | Algorithmacy is the focal construct. Antecedents are task interdependence and tenure. Moderators are perceived system authority and rule-change frequency. Outcomes are coordination quality with the counterpart and dispute rate. A figure shows the nomological net. | `mohlmann2021`, `kellogg2020` | +| 6 | Work remaining | 400 | Item generation, expert review, a pilot, CFA against AI literacy and the APJHR algorithmic-competency scale, and a panel. States what is done (the instrument draft) and what is not (no data). | — | +| | **Text total** | **3,300** | | | +| | References | ~900 | About 25 entries | | + +**Track choice:** Track 17 if the paper leads with coordination; Track 4 if it leads with worker voice under algorithmic management. The author decides. Only one track is allowed. diff --git a/submissions/ecis2027_algorithmacy_is/PAPER.md b/submissions/ecis2027_algorithmacy_is/PAPER.md new file mode 100644 index 00000000..16da4ccc --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/PAPER.md @@ -0,0 +1,120 @@ +[DRAFTED by Claude, 26 Oct 2026, from OUTLINE.md and ARGUMENT.md. For the author's read-aloud. ECIS 2027 short research paper, Track 17. Eight pages all-inclusive in the ECIS template; anonymized.] + +# ALGORITHMACY: A CONSTRUCT FOR COORDINATING THROUGH AN EVALUATING SYSTEM + +*Short Paper* + +## Abstract + +IS research has several constructs for how people work with algorithms: AI literacy, algorithm literacy, algorithmic competency and algorithm sensemaking. Each places the system in one position. It is a partner to one user, a channel between two users, or a manager over one worker. Many platforms hold a fourth position: they stand between two interdependent people, evaluate both, and bind both to their verdicts. This short paper defines algorithmacy as the competence to coordinate with another person through such a system. It has three facets. The first is interpreting an ambiguous result, the second is specifying intent so the counterpart can read it through the system, and the third is tracking rule changes that go unannounced. I set out a research model that separates algorithmacy from its rivals, and state the scale development and validation work that remains. + +**Keywords:** Algorithmacy, Algorithmic Management, AI Literacy, Coordination, Construct Development. + +## 1 Introduction + +A marketplace seller sees her search ranking fall after a buyer leaves a lukewarm review. She cannot tell what moved the ranking. It may have been the buyer's words, the platform's weighting of them, a change in the weighting itself, or her own conduct at some earlier point that the platform recorded and the buyer never saw. Curchod et al. (2020) studied sellers in exactly this position on eBay. Customer reviews there do not pass between buyer and seller as private messages. They enter the platform's tally, which judges the seller and decides her visibility. The relation is no longer a dyad between buyer and seller but a triad among operator, buyer and seller. + +What competence does the seller need? IS research offers several answers. Each answers a different question, because each assumes a different position for the system. AI literacy asks what a person should understand about a system they use. Algorithm literacy asks what a person knows about the algorithms that curate what they see. Algorithmic competency asks how a worker understands and works with the algorithms that assign and evaluate their work. Studies of AI-mediated communication ask how a system that modifies messages changes what two communicators think of each other. None asks what the seller needs: the competence to coordinate with another person through a system that judges them both. + +This paper argues that the gap is structural, not a matter of missing items. The existing constructs cannot be extended to cover the seller by adding questions, because they place the system in a different position relative to the people involved. I define algorithmacy as the competence that the fourth position demands. The construct is defined by what the evaluating system takes away from the two parties, so its facets follow from the structure rather than from a list of desirable skills. The paper then sets out discriminant logic against the two nearest rivals and a research model for testing the construct, and closes with the work that remains. + +## 2 Four Positions for the System + +Constructs for working with algorithms differ less in their items than in where they place the system relative to the people who use it. Table 1 sets out four positions. + +| Position of the system | Construct | Who is counted | What it leaves out | +|---|---|---|---| +| Partner to one user | AI literacy; algorithm literacy | One person and the system | A second person on the other side | +| Channel between two users | AI-mediated communication competence | Two people; the system carries messages | The system's own verdict | +| Manager over one worker | Algorithmic competency; algorithm sensemaking | One worker and the system | The client or counterpart | +| Evaluator between two people | Algorithmacy (this paper) | Two people and the system | — | + +*Table 1. Four positions for the system.* + +**Partner to one user.** Long and Magerko (2020) define AI literacy as a set of competencies that let an individual evaluate AI critically, communicate and collaborate with it, and use it as a tool. Their seventeen competencies are about what a person should understand of AI. One competency names other people, the humans who program and tune AI systems, but no competency names a counterpart with whom the user coordinates through the system. Algorithm literacy follows the same pattern. Dogruel et al. (2022) define it through awareness of algorithms, knowledge of how they work, critical evaluation of their decisions, and the skill to cope with or influence them. Two integrative reviews find a field with many overlapping constructs and no cohesive framework (Gagrčin et al., 2026; Oeldorf-Hirsch and Neubaum, 2025). In both, the relation studied is between a user and the platform that curates their feed. + +**Channel between two users.** Hancock et al. (2020, p. 89) define AI-mediated communication as "interpersonal communication in which an intelligent agent operates on behalf of a communicator by modifying, augmenting, or generating messages to accomplish communication goals." Here two people are counted. The system acts for one of them and changes what passes between them, and research asks how its presence changes what each thinks of the other. The system does not judge either of them. + +**Manager over one worker.** Studies of platform work place the system above the worker. Jarrahi and Sutherland (2019) describe workers who develop algorithmic competencies, a practical literacy for understanding and working with the algorithms that assign and assess their work. Sutherland et al. (2020, p. 470) name a "platform literacy, which is the know-how required to leverage platform resources in order to minimize the precarity of independent work." Möhlmann et al. (2023) describe algorithm sensemaking, the process by which workers make sense of algorithmic management. Zhou et al. (2025) built and validated the first scale of algorithmic competency, with four dimensions: understanding, embracing, leveraging and remediating algorithmic management. These constructs take the system's power seriously. They count one worker and one system. Two of Zhou et al.'s twelve items mention customers, but as objects of the worker's strategy toward the system. Item 8 reads, "I proactively explore AM rules to minimize negative customer feedback". No item asks about coordinating with the customer. + +**Evaluator between two people.** The fourth position is the seller's. Stark and Pais (2020, p. 53) describe the platform relation as "typically three-sided": the owner, the provider and the user-customer. The platform does not merely carry messages between provider and customer, and it does not merely manage the provider. It enrols the customer in evaluating the provider, fuses that evaluation with its own rule, and binds the provider to the result (Curchod et al., 2020). Möhlmann et al. (2021) show that on such platforms the algorithm matches the two sides and controls the provider at once. In this position the system is an evaluator between two people, and no existing construct counts all three. + +## 3 Algorithmacy Defined + +I define algorithmacy as the competence to coordinate with another person through a system that evaluates both and binds both to its verdict. The definition does not list skills. It names a structural position and asks what the position takes away from the two people in it. + +Coordinating directly, two people can ask each other what happened, say what they meant, and agree on the rules of their exchange. The evaluating system removes each of these. First, it fuses the counterpart's input with its own rule into one result, so neither party can tell who caused an outcome. Rahman (2021) found that freelancers on one platform could not tell how much of their score came from the client and how much from the platform. Second, every message to the counterpart also reaches the evaluator, so a message must be written for two readers. Third, the system changes its rule without notice, so the parties must infer change from shifts in their results. DeVito (2021) found that users of changing platforms keep revising their theories of how the system works, because the system moves under them. + +Each removal defines one facet of algorithmacy. + +- **Interpreting** is reading an ambiguous result to tell the counterpart's contribution from the system's. It restores the ability to attribute responsibility. +- **Specifying intent** is writing so that the counterpart can read what an action meant even as the system scores it. It restores common understanding between the parties. +- **Tracking** is noticing when the system's rule has changed, as distinct from a change in the counterpart. It restores predictability. + +The three facets share an object that no rival construct has: the relation between the two people, as the system shapes it. Interpreting asks who caused a result between us. Specifying intent asks what my action means to you. Tracking asks whether the rules between us have changed. A person can know a great deal about machine learning and still fail at all three, and a person with no technical knowledge can be good at all three through practice. + +## 4 Discriminant Logic + +A new construct earns its place only if it predicts something its rivals cannot. Algorithmacy should diverge from its two nearest rivals in identifiable cases. + +**Against AI literacy and algorithm literacy.** These measure what a person knows about algorithms. Algorithmacy should diverge from them where knowledge is high but the counterpart is ignored. A technically sophisticated seller may understand ranking systems well and still misread a drop in visibility as the buyer's fault when the platform changed its weighting, or the reverse. Knowledge of the system does not tell her which of the two acted. Algorithmacy predicts better attribution; literacy does not. + +**Against algorithmic competency.** Algorithmic competency measures a worker's strategy toward the system. Algorithmacy should diverge from it where a worker games the system but loses the counterpart. Bucher et al. (2021) describe workers who pacify the algorithm by curtailing their contact with clients to stay clear of its scrutiny. Such workers may score high on leveraging the system and still damage the relation the work depends on. Algorithmacy predicts that these workers coordinate worse with their counterparts; algorithmic competency predicts the opposite. + +The two divergences give the construct a testable boundary. In settings where the system holds the fourth position, algorithmacy should explain variance in coordination with the counterpart beyond what literacy and algorithmic competency explain. In settings where the system is only a partner, a channel or a manager, the three facets should add little. + +## 5 Research Model + +Figure 1 sets out the model. Algorithmacy is the focal construct, measured as three correlated facets. + +*Figure 1. Research model. Task interdependence → Algorithmacy (interpreting, specifying intent, tracking) → coordination quality with the counterpart; dispute rate. Moderators on the algorithmacy–outcome paths: perceived system authority; rule-change frequency. Controls: AI literacy; algorithmic competency.* + +**Antecedents.** Task interdependence should raise algorithmacy, because people whose outcomes depend on a counterpart through the system get more practice in reading and addressing both. Tenure on the system should raise it for the same reason. + +**Outcomes.** Algorithmacy should raise the quality of coordination with the counterpart: fewer misattributed failures, fewer misread messages and faster adjustment after rule changes. It should also lower the rate of disputes that arise from misattribution. + +**Moderators.** The effect of algorithmacy on outcomes should be stronger where the system's verdicts carry more authority over the parties, because misreading a binding verdict costs more than misreading an advisory one. It should also be stronger where the system changes its rules often, because tracking matters most when the rules move. Kellogg et al. (2020) describe algorithmic control as more encompassing, instantaneous, interactive and opaque than earlier forms, which is the condition under which these moderators should be high. + +**Controls.** AI literacy and algorithmic competency enter as controls. The discriminant claim of Section 4 is that algorithmacy predicts the outcomes over and above both. + +## 6 Work Remaining + +This short paper offers the construct and the model; the empirical work is ahead. Four steps remain before ECIS 2027 and after it. First, I am generating items for each facet from the structural definition, so that every item names the counterpart and the system together. Second, expert review will check that each item belongs to one facet and that no item could be answered without reference to a counterpart. Third, a pilot with platform sellers and hosts will test the three-facet structure by confirmatory factor analysis, with AI literacy and the Zhou et al. (2025) algorithmic-competency scale as discriminant measures. Fourth, a panel study will test the antecedent, outcome and moderator paths in Figure 1. By the conference I expect to report the item pool, the expert review and the pilot's factor structure. + +## 7 Conclusion + +The IS constructs for working with algorithms place the system as a partner, a channel or a manager. Many platforms place it between two people as an evaluator, and in that position the competence that matters is coordinating with the other person through the system. Algorithmacy names that competence. Its three facets follow from what the evaluator removes, and its discriminant predictions mark where it should and should not add to what IS research already measures. + +## References + +Bucher, E. L., Schou, P. K., and Waldkirch, M. (2021). "Pacifying the algorithm: Anticipatory compliance in the face of algorithmic management in the gig economy." *Organization* 28 (1), 44–67. + +Curchod, C., Patriotta, G., Cohen, L., and Neysen, N. (2020). "Working for an algorithm: Power asymmetries and agency in online work settings." *Administrative Science Quarterly* 65 (3), 644–676. + +DeVito, M. A. (2021). "Adaptive folk theorization as a path to algorithmic literacy on changing platforms." *Proceedings of the ACM on Human-Computer Interaction* 5 (CSCW2), Article 339. + +Dogruel, L., Masur, P., and Joeckel, S. (2022). "Development and validation of an algorithm literacy scale for internet users." *Communication Methods and Measures* 16 (2), 115–133. + +Gagrčin, E., Naab, T. K., and Grub, M. F. (2026). "Algorithmic media use and algorithm literacy: An integrative literature review." *New Media & Society* 28 (1), 423–447. + +Hancock, J. T., Naaman, M., and Levy, K. (2020). "AI-mediated communication: Definition, research agenda, and ethical considerations." *Journal of Computer-Mediated Communication* 25 (1), 89–100. + +Jarrahi, M. H. and Sutherland, W. (2019). "Algorithmic management and algorithmic competencies: Understanding and appropriating algorithms in gig work." In: *Information in Contemporary Society: iConference 2019 Proceedings*. Ed. by N. G. Taylor, C. Christian-Lamb, M. H. Martin, and B. Nardi. Lecture Notes in Computer Science 11420. Cham: Springer, 578–589. + +Kellogg, K. C., Valentine, M. A., and Christin, A. (2020). "Algorithms at work: The new contested terrain of control." *Academy of Management Annals* 14 (1), 366–410. + +Long, D. and Magerko, B. (2020). "What is AI literacy? Competencies and design considerations." In: *Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems*. New York: ACM, 1–16. + +Möhlmann, M., Alves de Lima Salge, C., and Marabelli, M. (2023). "Algorithm sensemaking: How platform workers make sense of algorithmic management." *Journal of the Association for Information Systems* 24 (1), 35–64. + +Möhlmann, M., Zalmanson, L., Henfridsson, O., and Gregory, R. W. (2021). "Algorithmic management of work on online labor platforms: When matching meets control." *MIS Quarterly* 45 (4), 1999–2022. + +Oeldorf-Hirsch, A. and Neubaum, G. (2025). "What do we know about algorithmic literacy? The status quo and a research agenda for a growing field." *New Media & Society* 27 (2), 681–701. + +Rahman, H. A. (2021). "The invisible cage: Workers' reactivity to opaque algorithmic evaluations." *Administrative Science Quarterly* 66 (4), 945–988. + +Stark, D. and Pais, I. (2020). "Algorithmic management in the platform economy." *Sociologica* 14 (3), 47–72. + +Sutherland, W., Jarrahi, M. H., Dunn, M., and Nelson, S. B. (2020). "Work precarity and gig literacies in online freelancing." *Work, Employment and Society* 34 (3), 457–475. + +Zhou, L., Lei, X., Liu, M., Huang, X., and Hou, R. (2025). "Algorithmic competency of on-demand labor platform workers: Scale development, antecedents, and consequences." *Asia Pacific Journal of Human Resources* 63 (2), e70004. diff --git a/submissions/ecis2027_algorithmacy_is/PAPER_SUBMITTED.md b/submissions/ecis2027_algorithmacy_is/PAPER_SUBMITTED.md new file mode 100644 index 00000000..9a8b6a28 --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/PAPER_SUBMITTED.md @@ -0,0 +1,133 @@ +[Author's edited version, from the Google Doc of 26 Oct 2026. This is the text built into the submission files.] + +# ALGORITHMACY: A CONSTRUCT FOR COORDINATING + +*Short Paper* + +## Abstract + +Although information systems research investigates human-algorithm interaction through constructs such as AI literacy, algorithmic competency, and algorithm sensemaking, prevailing models typically conceptualize the algorithm as a personal assistant, a passive transmission channel, or an automated supervisor. However, digital platforms increasingly position the algorithm directly between interdependent actors, evaluating both parties and enforcing binding outcomes. This paper introduces *algorithmacy*, defined as the capacity to coordinate with an interdependent counterpart through an evaluative platform. This competence entails deciphering ambiguous algorithmic feedback, structuring communicative intent so that a counterpart can interpret it through the system, and tracking unannounced shifts in platform governance. We contrast algorithmacy with adjacent constructs, present a research model linking its antecedents and coordination outcomes, and outline the methodological roadmap for scale development and empirical validation. + +**Keywords:** Algorithmacy, Algorithmic Management, AI Literacy, Coordination, Construct Development. + +## 1. Introduction + +When an online seller’s search rank drops after a lukewarm customer review, diagnosing the root cause is often intractable. The platform may have penalized specific phrasing in the review, recalibrated its sentiment weights, updated its ranking model without notice, or factored in historical account metrics invisible to the buyer. As Curchod et al. (2020) observed among eBay merchants, reviews do not operate as private bilateral feedback between trading partners. Instead, an algorithmic engine processes them to determine market visibility, transforming a dyadic transaction into a triadic relationship among buyer, seller, and platform operator. + +Current information systems (IS) research does not adequately capture the competence required in this environment because prevailing constructs frame algorithmic systems in fundamentally different roles. AI literacy examines how individuals operate computational tools; algorithm literacy considers how consumers interpret curated feeds; algorithmic competency focuses on how gig workers navigate top-down algorithmic management; and AI-mediated communication investigates how automated phrasing alters interpersonal perception. None of these constructs addresses the challenge of two interdependent parties coordinating through an opaque intermediary that simultaneously evaluates both sides and enforces binding outcomes. + +This mismatch is theoretical rather than psychometric: it cannot be resolved simply by appending new items to existing instruments. Existing scales assume either bilateral or hierarchical relationships, whereas algorithmacy emerges specifically from the demands of this triadic position. Because the construct reflects what an evaluative platform removes from direct interpersonal coordination, its core dimensions stem from structural platform mechanics rather than generic digital skill taxonomies. This paper establishes algorithmacy's discriminant boundaries, presents a research model for empirical testing, and outlines priorities for scale development and validation. + +## 2. Four Positions for the System + +The primary distinction among constructs governing human-algorithm interaction lies in how they configure the structural relationship between the system and its human users. Table 1 summarizes these configurations across four relational positions. + +*Table 1. Four Structural Positions of the System in Human-Algorithm Interaction* + +| Position of the System | Associated Constructs | Relational Scope | Excluded Dynamics | +| --- | --- | --- | --- | +| Partner to one user | AI literacy; algorithm literacy | Single user and the system | An interdependent counterpart on the other side | +| Channel between two users | AI-mediated communication competence | Two people; system acts as a conduit | The system’s independent evaluative verdict | +| Manager over one worker | Algorithmic competency; algorithm sensemaking | Single worker and the system | The customer or counterpart | +| Evaluator between two people | Algorithmacy (this paper) | Two interdependent users and the system | None | + +- **Partner to one user.** Under this framing, scholarship examines dyadic interactions between an isolated individual and a technology. For example, Long and Magerko’s (2020) conceptualization of AI literacy spans seventeen competencies detailing how an individual critically evaluates, collaborates with, and uses AI as a tool. While one competency acknowledges the professionals who build and tune these models, none accounts for an interdependent human counterpart with whom the user must coordinate. Algorithm literacy operates within the same boundary. Dogruel et al. (2022) define it in terms of user awareness, procedural knowledge, critical evaluation of system outputs, and individual coping strategies. As recent reviews emphasize (Gagrčin et al., 2026; Oeldorf-Hirsch & Neubaum, 2025), the literature remains fragmented across overlapping terms, yet almost uniformly centers on a lone user interpreting a curated feed. +- **Channel between two users.** This configuration brings interpersonal interaction into view, but sidelines the system’s independent evaluative authority. Hancock et al. (2020, p. 89) define AI-mediated communication as interpersonal exchange in which an intelligent agent modifies, augments, or generates messages on a communicator’s behalf to achieve communication goals. While this model accounts for two human actors, the algorithm operates primarily as a proxy or communicative aid. Research in this stream examines how automated linguistic adjustments shape mutual impressions, interpersonal trust, and relational dynamics; however, the system issues no binding administrative or economic judgments on either party. +- **Manager over one worker.** Scholarship on platform labor adopts a hierarchical perspective, positioning the system directly above the worker. Researchers examine how workers navigate computational control: developing practical tactics to decode automated assignments (Jarrahi & Sutherland, 2019), building platform literacy to buffer against structural precarity (Sutherland et al., 2020, p. 470), or engaging in algorithm sensemaking to decipher opaque supervisory practices (Möhlmann et al., 2023). Zhou et al. (2025) formalized this perspective by validating a four-dimensional scale of algorithmic competency—spanning understanding, embracing, leveraging, and remediating algorithmic management. Although these models take algorithmic authority seriously, their relational scope remains dyadic: an isolated worker negotiating with an automated supervisor. When counterparts appear, they are treated merely as external risk factors; Zhou et al.’s Item 8, for example, captures exploratory tactics to avoid negative customer feedback rather than strategies to coordinate with the customer. +- **Evaluator between two people.** The fourth position reflects the triadic reality of platform markets. As Stark and Pais (2020, p. 53) argue, platform organization is inherently three-sided, linking the platform owner, the service provider, and the customer. Rather than simply routing communication or exercising ordinary managerial oversight, the platform enlists one party to evaluate the other, synthesizes that input through proprietary algorithms, and enforces the outcome on both (Curchod et al., 2020). By simultaneously matching market participants and automatically controlling provider conduct (Möhlmann et al., 2021), the algorithm operates as an evaluative arbiter between interdependent individuals. No existing literacy or competency construct accounts for all three nodes of this triad. + +## 3. Algorithmacy Defined + +Algorithmacy is defined as *the competence to coordinate with another person through a system that evaluates both parties and enforces its decisions on each*. Rather than compiling an inventory of generic technical or cognitive skills, this definition begins from a structural position and identifies what that position strips away from direct interpersonal exchange. + +In direct, unmediated interaction, two people can ask what happened, clarify what they meant, and mutually negotiate the ground rules of their engagement. An evaluative platform dismantles each of these baseline coordination mechanisms: + +1. **Causal conflation:** The system aggregates counterpart input with internal scoring algorithms into a single score or decision, preventing both parties from isolating the true source of an outcome. For instance, Rahman (2021) found that freelance workers could not separate client feedback from automated platform adjustments when evaluating rating declines. +2. **Dual-audience communication:** Because every interaction passes through an automated monitor, messages must simultaneously address two distinct audiences: the human counterpart and the algorithmic scoring system recording their exchange. +3. **Unannounced rule volatility:** Platforms routinely adjust their evaluative and ranking logic without notice, forcing participants to infer institutional shifts from sudden variances in operational outcomes. As DeVito (2021) observed, users under volatile algorithmic regimes must continually revise their "folk theories" as the ground shifts beneath them. + +Each constraint corresponds directly to a constitutive facet of algorithmacy: + +- **Interpreting:** The capacity to parse ambiguous platform feedback and distinguish counterpart actions from system intervention, thereby restoring causal attribution. +- **Specifying intent:** The ability to formulate actions and communications so that a counterpart can understand them even while an automated evaluator monitors and scores the exchange, thereby preserving mutual understanding. +- **Tracking:** The discipline of detecting unannounced rule changes as distinct from shifts in counterpart behavior, thereby re-establishing predictability in exchange. + +These three facets share an analytical focus absent from adjacent constructs: how an evaluative intermediary reconfigures the relationship between two individuals. *Interpreting* resolves which party produced an outcome; *specifying intent* ensures that one’s meaning reaches the counterpart despite algorithmic mediation; and *tracking* determines whether the governing rules between the parties have shifted. + +Because algorithmacy centers on triadic coordination rather than computational knowledge, technical expertise is neither necessary nor sufficient. A software engineer well-versed in machine learning may struggle to coordinate effectively under these conditions, whereas a non-technical seller or gig worker may exhibit high algorithmacy honed through sustained platform practice. + +## 4. Discriminant Logic + +A construct achieves theoretical legitimacy by demonstrating distinct conceptual boundaries and incremental predictive utility relative to established alternatives. For algorithmacy, this boundary becomes evident when technical sophistication or platform-gaming tactics fail to produce effective interpersonal coordination. + +- **Distinction from AI and algorithm literacy:** AI literacy and algorithm literacy assess an individual’s technical and conceptual grasp of computational systems. Algorithmacy diverges from both whenever technical understanding is high but relational attunement to the counterpart is absent. A merchant may understand neural networks, collaborative filtering, and ranking models, yet still misattribute a sudden loss of search visibility to a buyer’s review when the platform actually revised its internal weighting formula—or vice versa. Knowing how algorithms work in the abstract does not clarify which party generated a specific outcome. Algorithmacy accounts for this relational attribution; technical literacy does not. +- **Distinction from algorithmic competency:** Algorithmic competency measures how workers understand, leverage, and resist automated supervision. Algorithmacy diverges from this construct when optimizing for the system alienates the human counterpart. Bucher et al. (2021), for example, describe platform workers who placate supervisory algorithms by deliberately curtailing communication with clients to avoid platform flags. A worker who excels at manipulating platform metrics this way may score high on algorithmic competency while actively undermining the collaborative relationship the work depends on. Algorithmic competency treats these evasion tactics as mastery; algorithmacy identifies them as coordination failures. + +These conceptual differences establish clear empirical boundaries. Where a system acts as an evaluative arbiter between two interdependent actors, algorithmacy should explain variance in coordination performance beyond what general literacy and supervisory competency scales capture. Conversely, where technology functions merely as a personal tool, a passive communication conduit, or a top-down supervisor, the three facets of algorithmacy should offer little incremental explanatory power. + +## 5. Research Model and Hypotheses + +FIGURE1 + +Because algorithmacy develops through situated practice rather than formal instruction, its emergence depends on an individual's immersion in triadic coordination. Two antecedents are primary: + +- **Task interdependence:** When participants depend heavily on a counterpart to achieve their goals, every transaction compels them to determine whether friction stems from partner behavior or platform intervention. +- **Platform tenure:** Accumulated experience under an evaluative system provides the empirical feedback loops required to separate counterpart idiosyncrasies from systemic algorithmic patterns. + +In turn, algorithmacy directly improves the quality of interpersonal coordination. Actors who accurately parse ambiguous feedback and signal intent across an automated monitor will experience fewer misattributed failures, encounter fewer communicative misunderstandings, and adapt more swiftly to unannounced system adjustments. By clarifying intent and decoupling partner conduct from platform interference, algorithmacy should also significantly reduce disputes stemming from misread feedback. + +Two structural moderators bound these coordination benefits: + +- **Evaluative authority:** The performance dividend of algorithmacy depends on the degree of control the platform exercises over the actors. When algorithmic verdicts carry binding consequences—such as search demotions, automated penalty fees, or account deactivations—misreading feedback is immediately and severely costly. Conversely, where ratings lack binding enforcement or remain advisory, coordination errors carry lower stakes. +- **Algorithmic volatility:** Tracking competencies become salient primarily in unstable operating environments. Frequent, unannounced rule changes heighten the risk of coordinating on obsolete assumptions. As Kellogg et al. (2020) observe, contemporary algorithmic control is encompassing, instantaneous, interactive, and opaque. Where platforms exhibit these characteristics most acutely, continuously tracking shifting algorithmic rules is essential to sustaining exchange. + +Empirical testing will control for AI literacy and algorithmic competency. Establishing discriminant validity requires showing that algorithmacy explains significant incremental variance in coordination quality and dispute reduction beyond general technical knowledge and individual gaming tactics. + +## 6. Work Remaining + +While this paper formalizes the conceptual foundations and structural model of algorithmacy, the empirical validation program proceeds across four sequential phases: + +1. **Item generation:** Draft indicators for each facet are being derived directly from the triadic framework. Rather than assessing generic technical tasks, every item anchors on the relational exchange, capturing how the system intervenes between the user and their counterpart. +2. **Content validation:** An expert panel will evaluate the initial item pool to establish content validity, confirm facet mapping, and verify that no indicator can be answered without accounting for an interdependent counterpart. +3. **Factor structure and discriminant testing:** A pilot study of online marketplace sellers and platform hosts will evaluate the proposed three-factor model via confirmatory factor analysis (CFA) and test discriminant validity against established scales of AI literacy and Zhou et al.’s (2025) algorithmic competency instrument. +4. **Model testing:** A longitudinal panel study will test the hypothesized antecedents, coordination outcomes, and structural moderators specified in the research model. + +By ECIS 2027, we expect to report the refined item pool, expert panel assessments, and preliminary factor structure from the pilot administration. + +## 7. Conclusion + +Most information systems constructs frame human interaction with algorithms through familiar relational paradigms: the system as an individual’s tool, a conduit between communicators, or an automated supervisor. Modern platforms, however, increasingly occupy a fourth position—standing directly between interdependent individuals as an evaluative arbiter that enforces binding outcomes on both. In these environments, technical literacy and metric gaming are insufficient; what matters is the capacity to coordinate with another human being through an automated monitor. Algorithmacy captures this competence, anchors its core dimensions in platforms' structural constraints, and provides a precise theoretical foundation for future empirical inquiry. + +## References + +Bucher, E. L., Schou, P. K., & Waldkirch, M. (2021). “Pacifying the algorithm: Anticipatory compliance in the face of algorithmic management in the gig economy.” *Organization* 28 (1), 44–67. + +Curchod, C., Patriotta, G., Cohen, L., and Neysen, N. (2020). “Working for an algorithm: Power asymmetries and agency in online work settings.” *Administrative Science Quarterly* 65 (3), 644–676. + +DeVito, M. A. (2021). “Adaptive folk theorization as a path to algorithmic literacy on changing platforms.” *Proceedings of the ACM on Human-Computer Interaction* 5 (CSCW2), Article 339. + +Dogruel, L., Masur, P., and Joeckel, S. (2022). “Development and validation of an algorithm literacy scale for internet users.” *Communication Methods and Measures* 16 (2), 115–133. + +Gagrčin, E., Naab, T. K., and Grub, M. F. (2026). “Algorithmic media use and algorithm literacy: An integrative literature review.” *New Media & Society* 28 (1), 423–447. + +Hancock, J. T., Naaman, M., and Levy, K. (2020). “AI-mediated communication: Definition, research agenda, and ethical considerations.” *Journal of Computer-Mediated Communication* 25 (1), 89–100. + +Jarrahi, M. H. and Sutherland, W. (2019). “Algorithmic management and algorithmic competencies: Understanding and appropriating algorithms in gig work.” In: *Information in Contemporary Society: iConference 2019 Proceedings*. Ed. by N. G. Taylor, C. Christian-Lamb, M. H. Martin, and B. Nardi. Lecture Notes in Computer Science 11420. Cham: Springer, 578–589. + +Kellogg, K. C., Valentine, M. A., and Christin, A. (2020). “Algorithms at work: The new contested terrain of control.” *Academy of Management Annals* 14 (1), 366–410. + +Long, D. and Magerko, B. (2020). “What is AI literacy? Competencies and design considerations.” In: *Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems*. New York: ACM, 1–16. + +Möhlmann, M., Alves de Lima Salge, C., and Marabelli, M. (2023). “Algorithm sensemaking: How platform workers make sense of algorithmic management.” *Journal of the Association for Information Systems* 24 (1), 35–64. + +Möhlmann, M., Zalmanson, L., Henfridsson, O., and Gregory, R. W. (2021). “Algorithmic management of work on online labor platforms: When matching meets control.” *MIS Quarterly* 45 (4), 1999–2022. + +Oeldorf-Hirsch, A. and Neubaum, G. (2025). “What do we know about algorithmic literacy? The status quo and a research agenda for a growing field.” *New Media & Society* 27 (2), 681–701. + +Rahman, H. A. (2021). “The invisible cage: Workers’ reactivity to opaque algorithmic evaluations.” *Administrative Science Quarterly* 66 (4), 945–988. + +Stark, D. and Pais, I. (2020). “Algorithmic management in the platform economy.” *Sociologica* 14 (3), 47–72. + +Sutherland, W., Jarrahi, M. H., Dunn, M., and Nelson, S. B. (2020). “Work precarity and gig literacies in online freelancing.” *Work, Employment and Society* 34 (3), 457–475. + +Zhou, L., Lei, X., Liu, M., Huang, X., and Hou, R. (2025). “Algorithmic competency of on-demand labor platform workers: Scale development, antecedents, and consequences.” *Asia Pacific Journal of Human Resources* 63 (2), e70004. diff --git a/submissions/ecis2027_algorithmacy_is/README.md b/submissions/ecis2027_algorithmacy_is/README.md new file mode 100644 index 00000000..c6aff57d --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/README.md @@ -0,0 +1,29 @@ +# ECIS 2027: Algorithmacy as an IS construct + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. The angle is a proposal for the author to rule on.] + +**Status: SUBMITTED 26 Oct 2026** (PCS submission 1520, Track 17, Short Paper). The submitted text is `PAPER_SUBMITTED.md` (the author's edit of `PAPER.md`). Acceptance is due 1 Feb 2027. + +**The call:** European Conference on Information Systems, Luxembourg, May 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Short research paper (8 pages, all-inclusive) | **30 Oct 2026** | +| Acceptance | 1 Feb 2027 | +| Conference | May 2027 | + +**Angle:** existing constructs place the system as partner, channel or manager. Algorithmacy is defined by a fourth position: an evaluator between two people. The paper offers a research model and discriminant logic. + +**Relation to other arms:** +- Spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/). The measurement companion is [`../perspsych_measuring_algorithmacy/`](../perspsych_measuring_algorithmacy/). +- The instrument draft lives in `org_frontier/survey/cohort_algorithmacy/`. + +| File | What it is | +|---|---| +| `ABSTRACT.md` | Title and 142-word abstract, with its citation check | +| `ARGUMENT.md` | P1, P2, C | +| `OUTLINE.md` | 6-section outline sized to 8 pages | +| `FLAGS.md` | Concurrency, evidence and fit cautions | +| `cfp/` | Digest; `source/` is gitignored | + +**Guardrails:** Claude does not submit. The author's read-aloud is the last gate. diff --git a/submissions/ecis2027_algorithmacy_is/cfp/README.md b/submissions/ecis2027_algorithmacy_is/cfp/README.md new file mode 100644 index 00000000..b3771a46 --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/cfp/README.md @@ -0,0 +1,31 @@ +# CFP brief: ECIS 2027 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://ecis2027.uni.lu/?p=3130 (Call for Submissions) +- https://ecis2027.uni.lu/track-04/ +- https://ecis2027.uni.lu/track-17/ +- https://ecis2027.uni.lu/important-dates/ + +## Dates (verbatim) +- "Paper submissions: 30 October 2026". No timezone is stated. +- "Paper acceptance: 1 February 2027". The conference is in May 2027, in Luxembourg. + +## Format (verbatim) +- "Short Research Papers must not exceed eight (8) single-spaced pages ... The page limit is all inclusive, encompassing the title, body of the paper, figures and tables, references, and appendices." +- Short papers "can be theoretical (presenting promising frameworks)". They "should include a paragraph describing what work remains to complete the paper". +- Submit through PCS using the ECIS template. "All papers must submitted to one – and only one – conference track." An author may appear on at most three submissions. +- Generative AI use must be declared in PCS. The AI reviewer is on by default, and authors can opt out. + +## Originality (verbatim) +- "A submission, or a similar version, must not have been published or accepted in a journal or another conference proceedings. Further, submissions must not be concurrently under consideration for publication or presentation elsewhere." +- "Accepted short papers ... will appear in the ECIS 2027 Proceedings." + +## Track lines nearest this paper +- **Track 4, Algorithmic management and decision-making** (Alizadeh, Tarafdar, Remus): + - "how workers retain agency and voice within algorithmically managed organisations" + - "How can participatory approaches to algorithmic management bridge tensions between workers, managers, and the systems that mediate their relationships?" +- **Track 17, AI and the future of organising** (Watkowski, Hönigsberg, O'Leary): + - "Human–AI collaboration and evolving forms of coordination, particularly in generative and agentic AI settings" + - "Sensemaking, framing, and legitimization of AI adoption and use" diff --git a/submissions/ecis2027_algorithmacy_is/figures/figure1.png b/submissions/ecis2027_algorithmacy_is/figures/figure1.png new file mode 100644 index 00000000..606eef10 Binary files /dev/null and b/submissions/ecis2027_algorithmacy_is/figures/figure1.png differ diff --git a/submissions/ecis2027_algorithmacy_is/figures/make_figure1.py b/submissions/ecis2027_algorithmacy_is/figures/make_figure1.py new file mode 100644 index 00000000..4b268fbe --- /dev/null +++ b/submissions/ecis2027_algorithmacy_is/figures/make_figure1.py @@ -0,0 +1,60 @@ +from PIL import Image, ImageDraw, ImageFont +import math +W,H=2010,1030 +im=Image.new('RGB',(W,H),'white'); d=ImageDraw.Draw(im) +F='/System/Library/Fonts/Supplemental/Arial.ttf'; FB='/System/Library/Fonts/Supplemental/Arial Bold.ttf'; FI='/System/Library/Fonts/Supplemental/Arial Italic.ttf' +f=ImageFont.truetype(F,34); fb=ImageFont.truetype(FB,36); fs=ImageFont.truetype(F,30); fi=ImageFont.truetype(FI,30); fsign=ImageFont.truetype(FB,38) +K=(0,0,0) +def box(x0,y0,x1,y1,lines,font=f,bold_first=False,dash=False,w=4): + if dash: + for (a,b,c,e) in [(x0,y0,x1,y0),(x1,y0,x1,y1),(x1,y1,x0,y1),(x0,y1,x0,y0)]: + L=math.hypot(c-a,e-b); n=int(L//24) + for i in range(0,n,2): + t0=i/n; t1=min((i+1)/n,1) + d.line([(a+(c-a)*t0,b+(e-b)*t0),(a+(c-a)*t1,b+(e-b)*t1)],fill=K,width=w) + else: d.rectangle([x0,y0,x1,y1],outline=K,width=w) + hs=[]; + for i,l in enumerate(lines): + ft=fb if (bold_first and i==0) else font + bb=d.textbbox((0,0),l,font=ft); hs.append((l,ft,bb[2]-bb[0],bb[3]-bb[1])) + tot=sum(h for *_,h in hs)+12*(len(hs)-1); y=(y0+y1)/2-tot/2 + for l,ft,w_,h in hs: + d.text(((x0+x1)/2-w_/2,y),l,font=ft,fill=K); y+=h+12 +def arrow(p,q,dash=False,w=4,head=True): + (x0,y0),(x1,y1)=p,q + if dash: + L=math.hypot(x1-x0,y1-y0); n=int(L//22) + for i in range(0,n-1,2): + t0=i/n;t1=(i+1)/n; d.line([(x0+(x1-x0)*t0,y0+(y1-y0)*t0),(x0+(x1-x0)*t1,y0+(y1-y0)*t1)],fill=K,width=w) + else: d.line([p,q],fill=K,width=w) + if not head: return + a=math.atan2(y1-y0,x1-x0); s=22 + d.polygon([(x1,y1),(x1-s*math.cos(a-0.4),y1-s*math.sin(a-0.4)),(x1-s*math.cos(a+0.4),y1-s*math.sin(a+0.4))],fill=K) +# antecedents +box(40,330,420,470,['Task','interdependence']) +box(40,610,420,750,['Tenure on','the system']) +# algorithmacy +box(640,300,1180,770,['Algorithmacy','','Interpreting','Specifying intent','Tracking'],bold_first=True,w=6) +# outcomes +box(1520,330,1940,470,['Coordination quality','with the counterpart']) +box(1520,610,1940,750,['Dispute rate']) +# moderators, between the two outcome paths +box(1250,480,1470,600,['Moderators:','system authority;','rule-change freq.'],font=ImageFont.truetype(F,24)) +# controls +box(1520,880,1940,990,['Controls: AI literacy,','algorithmic competency'],font=fs,dash=True) +# antecedent arrows +arrow((420,400),(640,460)); d.text((505,375),'+',font=fsign,fill=K) +arrow((420,680),(640,610)); d.text((505,655),'+',font=fsign,fill=K) +# outcome paths +arrow((1180,430),(1520,400)); d.text((1230,370),'+',font=fsign,fill=K) +arrow((1180,640),(1520,680)); d.text((1230,655),'−',font=fsign,fill=K) +# moderator arrows onto the paths +arrow((1360,480),(1360,418)); arrow((1360,600),(1360,658)) +# control arrows +arrow((1730,880),(1730,750),dash=True) +d.line([(1940,935),(1975,935)],fill=K,width=4) +arrow((1975,935),(1975,400),dash=True,head=False) +d.line([(1975,400),(1950,400)],fill=K,width=4) +d.polygon([(1940,400),(1962,389),(1962,411)],fill=K) +im=im.crop((0,270,W,1010)); im.save('/private/tmp/claude-501/-Users-ludwitt/b7816862-76fb-465e-84f5-2874bdfbfe95/scratchpad/ecis_build/figure1.png',dpi=(300,300)) +print('ok') diff --git a/submissions/egos2027_st64_rules_of_play/ABSTRACT.md b/submissions/egos2027_st64_rules_of_play/ABSTRACT.md new file mode 100644 index 00000000..e48b2af4 --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/ABSTRACT.md @@ -0,0 +1,24 @@ +# Abstract: EGOS 2027, sub-theme 64 (alternate to the sub-theme 03 short paper) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. EGOS allows one short paper per main author, so this is an alternate. Due 15 Jan 2027, 12:00 CET.] + +**Title:** Playing by Rules No One Wrote: Keeping the Counterpart in View Under Algorithmic Mediation + +## Abstract +The sub-theme asks who sets the rules of play in organizations mediated by opaque algorithms. On platforms, and increasingly inside firms, the answer is that a system sets them. Neither of the people it connects wrote those rules, and the system changes them without notice. A hotel and its guests, a seller and a buyer, or a manager and an employee whose appraisal an AI tool may draft, all coordinate through a system that evaluates both sides and binds both to its verdict. This paper asks what happens to human connection in that arrangement and what keeps it alive. + +The argument starts from Simmel's third party. A mediator helps two people settle their own affairs. An arbitrator takes the decision out of their hands. An opaque evaluating system is an arbitrator that neither party chose. Every message one person sends the other is also scored, and every result mixes the other person's judgement with the system's rule. People in this position lose each other in two ways. Some pacify the algorithm, curtailing contact with the counterpart to stay clear of scrutiny. Others treat the system as a neutral channel and read its scores as personal verdicts, missing the rule that judges them both. + +Two practices keep the counterpart in view. Specifying intent means writing so that the other person can read what an action meant through the system's filter. Tracking means noticing when the rule has changed, as distinct from a change in the other person. Together these practices are part of what the author calls algorithmacy: the competence to coordinate through an evaluating intermediary. The paper concludes that humanness under algorithmic mediation is not only resistance to the machine. It is also the practical work of reaching another person through a machine that judges both of them. + +**Keywords:** algorithmic mediation; humanness; third party; rules; coordination + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Mediator versus arbitrator | `triadic_reduction/library/cards/simmel1902number2.md` (p. 172) | page-image loci | +| No procedure to contest | `lima_pdw/literature/cards/kellogg2020.md` (p. 387) | full text | +| Curtailing contact to avoid scrutiny | `lima_pdw/literature/cards/bucherschouwaldkirch2021.md` | abstract (full text read in the OT claim check) | +| Double negotiation with guests and the algorithm | `dxd_foundation/literature/library/jhaver2018algorithmic.md` | full text | +| Hotel rankings | `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3) | full text | +| Reviews read as compliments (reciprocity) | — | needs a card for Bolton et al. 2013 (FLAGS 4) | diff --git a/submissions/egos2027_st64_rules_of_play/ARGUMENT.md b/submissions/egos2027_st64_rules_of_play/ARGUMENT.md new file mode 100644 index 00000000..cd83a555 --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/ARGUMENT.md @@ -0,0 +1,11 @@ +# Argument: Who sets the rules of play, and what keeps people connected through the machine + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** When an opaque algorithm mediates work between two people, it sets rules that neither of them wrote, and it changes them without notice. Its verdicts bind both people. +- *Sources:* `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14); `lima_pdw/literature/cards/kellogg2020.md` (p. 387: there is often "no procedure ... to contest"); `lima_pdw/literature/cards/rahman2021.md`. + +**P2.** Human connection at work depends on each person being able to tell what the other meant and did. Through such an algorithm, every message to the other person is also a message to the evaluator, and every result mixes the other person's judgement with the rule's. +- *Sources:* `lima_pdw/literature/cards/bucherschouwaldkirch2021.md` (anticipatory compliance; chat read as monitored); `dxd_foundation/literature/library/jhaver2018algorithmic.md` (hosts' "double negotiation" with guests and the algorithm). + +**C.** Connection survives algorithmic mediation through two practices that let people play by rules they did not set without losing each other. The first is specifying intent so the other person can read it through the system. The second is tracking the rules as they change. People lose connection in two ways. One is pacifying the algorithm and abandoning the counterpart. The other is treating the system as a neutral channel and missing the rule that is judging them both. diff --git a/submissions/egos2027_st64_rules_of_play/FLAGS.md b/submissions/egos2027_st64_rules_of_play/FLAGS.md new file mode 100644 index 00000000..9066cd9d --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/FLAGS.md @@ -0,0 +1,10 @@ +# Flags: EGOS 2027, sub-theme 64 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **One short paper per main author.** The author's Google Doc "Algorithmacy - EGOS Short Paper Submission" targets sub-theme 03 (Hyper-Real). Choose one. Sub-theme 64 asks the rules-of-play question directly; sub-theme 03 asks about simulations and data. This arm exists so the choice can be made on two drafts. +2. **The short-paper length is unconfirmed.** The guidelines PDF was unreadable. Read it in a browser before writing. +3. **"the author calls algorithmacy"** names the author in an anonymized short paper if EGOS reviews blind. Check the EGOS guidelines and change it to "we call" or "this paper calls" if needed. +4. **Bolton et al. (2013)** on reciprocal reviewing has no card. The abstract does not cite it, so the last citation-check row is a reminder for the full paper. +5. **Bucher, Schou & Waldkirch (2021)** is Eliane Bucher, not Taina Bucher. The lima card is abstract-only, but the OT claim check read the full text. +6. **Simmel's arbitrator is chosen by the parties.** The abstract says so ("that neither party chose" marks the difference). diff --git a/submissions/egos2027_st64_rules_of_play/OUTLINE.md b/submissions/egos2027_st64_rules_of_play/OUTLINE.md new file mode 100644 index 00000000..07d3353c --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/OUTLINE.md @@ -0,0 +1,12 @@ +# Outline: EGOS short paper (about 3,000 words; confirm the cap) and full paper (20–25 pages by mid-June 2027) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Short paper | Full paper | Mechanism | Cards | +|---|---|---|---|---|---| +| 1 | Who sets the rules | 400 | 1,000 | The system writes and rewrites rules that bind both parties. | `starkvandenbroeck2024`, `kellogg2020`, `rahman2021` (theory only) | +| 2 | The arbitrator nobody chose | 600 | 1,500 | Simmel's mediator–arbitrator contrast, with the difference that on platforms nobody chose the arbiter. | `simmel1902number2`, `simmel1950sociology` | +| 3 | Two ways to lose the counterpart | 700 | 2,000 | Pacifying the algorithm; treating it as a channel. | `bucherschouwaldkirch2021`, `jhaver2018algorithmic`, `eslami2019` | +| 4 | Two ways to keep the counterpart | 800 | 2,500 | Specifying intent and tracking change. Hotel and marketplace cases. | `scottorlikowski2012accountability`, `orlikowskiscott2023`, `curchod2020`, `devito2021` | +| 5 | Humanness as practical work | 500 | 1,500 | Answers the sub-theme's questions on humanness and rules. | — | +| | **Total** | **3,000** | **8,500** | | | diff --git a/submissions/egos2027_st64_rules_of_play/README.md b/submissions/egos2027_st64_rules_of_play/README.md new file mode 100644 index 00000000..85aee548 --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/README.md @@ -0,0 +1,17 @@ +# EGOS 2027, sub-theme 64: Playing by Rules No One Wrote + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. An alternate to the author's sub-theme 03 short paper.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **The short paper is due 15 Jan 2027, 12:00 CET.** Nothing has been submitted. Only one EGOS short paper per main author is allowed (FLAGS 1). + +**The call:** 43rd EGOS Colloquium, Liverpool, 8–10 July 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Short paper | **15 Jan 2027, 12:00 CET** | +| Decisions | end of February 2027 | +| Full paper (20–25 pages) | mid-June 2027 | + +**Angle:** answers the convenors' question "who sets the rules of play" with the triad. Human connection survives through specifying intent and tracking rules. + +**Relation to other arms:** an alternate to the sub-theme 03 draft (Google Doc). Builds on [`../algorithmacy_ot/`](../algorithmacy_ot/). diff --git a/submissions/egos2027_st64_rules_of_play/cfp/README.md b/submissions/egos2027_st64_rules_of_play/cfp/README.md new file mode 100644 index 00000000..b30597f8 --- /dev/null +++ b/submissions/egos2027_st64_rules_of_play/cfp/README.md @@ -0,0 +1,29 @@ +# CFP brief: EGOS 2027, sub-theme 64 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://www.egos.org/2027_Liverpool/CALL_for_short_papers +- Sub-theme 64 page: https://www.egos.org/jart/prj3/egos/main.jart?rel=de&reserve-mode=active&content-id=1787718887446&subtheme_id=1764791093250 +- The short-paper guidelines PDF and the sub-theme PDF came back as unreadable binary. + +**Event:** 43rd EGOS Colloquium, "The Games We Play: Re-Writing the Rules of Organization?", University of Liverpool Management School, 8–10 July 2027. + +## Dates (verbatim) +- "The submission window for short papers will open in early October 2026 and close on January 15, 2027 (12:00 CET)." +- "Decisions can be expected by the end of February 2027." +- "If your short paper is accepted, you are expected to submit a full paper version (approx. 20-25 pages) by middle of June, 2027 at the latest." + +## Limits (verbatim) +- **"You can only submit one paper to one sub-theme (i.e. main authorship)."** +- "You can be listed as co-author in a maximum of two additional papers." +- **The short-paper length is in the unreadable guidelines PDF. Confirm it in a browser.** EGOS short papers have usually been capped at 3,000 words, but that is not verified for 2027. + +## Sub-theme 64 +- **Title:** "Rage Against the Digital Machine? Rethinking Humanness and the Possibility for Human Connection in Times of Emerging Technologies" +- **Convenors:** Frank Martela (Aalto University), Sophia Galière (Université Côte d'Azur), Johan Alvehus (Lund University) +- **Lines from the call:** + - "Who sets the rules of play in organizations mediated by opaque algorithms? How do new digital tools reshape the balance of power and authority within the organization?" + - "How do workers, leaders, and algorithms co-construct new rules of the game—bending, hacking, or resisting the systems that seek to manage them?" + - "How do digital traces, status indicators, or AI analytics redefine what counts as presence, effort, or belonging?" + - "We welcome conceptual, empirical, and methodological contributions". diff --git a/submissions/euram2027_algorithmacy/ABSTRACT.md b/submissions/euram2027_algorithmacy/ABSTRACT.md new file mode 100644 index 00000000..f297a1c9 --- /dev/null +++ b/submissions/euram2027_algorithmacy/ABSTRACT.md @@ -0,0 +1,29 @@ +# Abstract: EURAM 2027, ST09_03 + +[DRAFTED by Claude, 26 Oct 2026. The author's own OT abstract (`../algorithmacy_ot/`) may be used instead; this version is written for an OB/HRM track. The limit is 400 words. Due 12 Jan 2027, 14:00 Brussels.] + +**Title:** Algorithmacy: The Competence to Coordinate When an Algorithm Judges Both Sides + +## Abstract +Organizations increasingly manage people through algorithms that evaluate their work, yet they rarely ask what competence that arrangement demands. Research offers several answers, and each treats the system as standing in one place. AI literacy treats the system as a partner to one user. Studies of computer-mediated communication treat it as a channel between two users. Work on algorithmic competency treats it as a manager over one worker. None of them covers a common case in algorithmically managed work: two people who depend on each other, such as a seller and a buyer, a hotel and its guests, or an employee and a manager reading an AI-generated appraisal, while a system evaluates both and binds both to its verdict. + +This paper classifies communicative competence by what stands between the people coordinating. In oracy they speak directly. In literacy a text moderates them, but a human still interprets and judges it. In what this paper calls algorithmacy, an opaque, adaptive system stands between them as an arbiter that neither party chose. Following Okhuysen and Bechky, coordination depends on accountability, predictability and common understanding. The arbiter dissolves all three. Neither party can tell whether an outcome came from the counterpart or from the rule, every message to the counterpart is also scored, and the rule changes without notice. + +Algorithmacy is the competence that restores those conditions, and it has three practices. Interpreting ambiguous results restores accountability. Specifying intent so the counterpart can read it through the system restores common understanding. Tracking rule changes the operator does not announce restores predictability. The paper describes two failure modes. In the first, a person collapses the triad into a pair with the machine and ignores the counterpart. In the second, a person treats the system as a neutral channel and ignores its judgement. + +For OB and HRM, the argument has three implications. First, simplifying a task through algorithmic management raises its relational demands, so deskilling and the demand for algorithmacy rise together. Second, people subject to the same rules will differ in outcomes according to this competence, which training in AI literacy does not supply. Third, if generative AI enters appraisal, hiring and collaboration, these demands will move from platform work into professional work. + +**Keywords:** algorithmic management; algorithmacy; coordination; competence; triads + +**Word count:** 368 (body, by whitespace split). + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| AI literacy as an individual's competencies | `lima_pdw/literature/cards/long2020.md` | full text | +| CMC competence | `lima_pdw/literature/cards/spitzberg2006.md` | full text | +| Algorithmic competency | `lima_pdw/literature/cards/jarrahisutherland2019.md`; `lima_pdw/literature/cards/zhou2025apjhr.md` | abstract; full text | +| Seller–buyer triad | `lima_pdw/literature/cards/curchod2020.md` | full text | +| Hotel rankings | `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3) | full text | +| Three integrating conditions | `lima_pdw/literature/cards/okhuysen2009.md` | full text | +| GenAI in professional work (as conjecture) | `lima_pdw/literature/cards/hindsvonkrogh2024.md` | full text | diff --git a/submissions/euram2027_algorithmacy/ARGUMENT.md b/submissions/euram2027_algorithmacy/ARGUMENT.md new file mode 100644 index 00000000..ef740f0b --- /dev/null +++ b/submissions/euram2027_algorithmacy/ARGUMENT.md @@ -0,0 +1,15 @@ +# Argument: the OT paper in short form, for ST09_03 + +[DRAFTED by Claude, 26 Oct 2026. This restates the OT paper's argument for a conference audience. The argument is the author's; the P/P/C framing is Claude's.] + +**P1.** Communicative competences are defined by what stands between the people coordinating. Oracy covers direct and chained dyads, and literacy covers dyads moderated by a text that a human still interprets and judges. +- *Sources:* the OT manuscript, §"What Stands Between as a Defining Criterion"; `lima_pdw/literature/cards/simmel1902.md` (group size and social form). + +**P2.** Algorithmic management places an opaque, adaptive intermediary between two people. It evaluates both, binds both to its verdict, and was chosen by neither. This turns the dyad into a triad and dissolves the integrating conditions of coordination: accountability, predictability and common understanding. +- *Sources:* + - `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 1, p. 14) + - `lima_pdw/literature/cards/curchod2020.md` + - `lima_pdw/literature/cards/burrell2016.md` (abstract depth) + - `lima_pdw/literature/cards/okhuysen2009.md` (p. 463 for the definition of coordination; the three conditions at pp. 483–490) + +**C.** Working in that triad requires a distinct competence, algorithmacy. It has three practices, each of which restores one condition: interpreting ambiguous results restores accountability, specifying intent across the filter restores common understanding, and tracking unannounced rule changes restores predictability. This gives HRM and OB a competence to develop, alongside AI literacy rather than inside it. diff --git a/submissions/euram2027_algorithmacy/FLAGS.md b/submissions/euram2027_algorithmacy/FLAGS.md new file mode 100644 index 00000000..cc077837 --- /dev/null +++ b/submissions/euram2027_algorithmacy/FLAGS.md @@ -0,0 +1,13 @@ +# Flags: EURAM 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **This is a full paper, due 12 Jan 2027 at 14:00 Brussels time.** The submission platform opens in late November. +2. **Concurrent journal review is not addressed.** EURAM bars papers "published or accepted for publication ... in any form – online preprint, working paper". It says nothing about papers under journal review. If the OT paper is accepted before June 2027, the EURAM version would breach the rule. Check *Organization Theory*'s policy on conference presentation too. +3. **No preprint or working paper of the OT paper may be posted** before the conference. +4. **One presenting author, one paper.** If the author presents at EURAM, no other EURAM paper. +5. **Track choice:** ST09_03 (job quality, autonomy, fairness) fits best. T09_04 explicitly welcomes conceptual work. Only one track is allowed. +6. **Reviewing duty:** submitting commits the author to review two EURAM papers. +7. **Example rule:** no driver or freelancer lead examples. +8. **The abstract is 368 words,** under the 400-word cap. +9. **Okhuysen & Bechky:** the three conditions are defined around pp. 483–490. Do not cite p. 463 for them. diff --git a/submissions/euram2027_algorithmacy/OUTLINE.md b/submissions/euram2027_algorithmacy/OUTLINE.md new file mode 100644 index 00000000..44eeaf27 --- /dev/null +++ b/submissions/euram2027_algorithmacy/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: EURAM 2027 full paper (10–40 pages) + +[DRAFTED by Claude, 26 Oct 2026. The OT manuscript (`../algorithmacy_ot/`, about 10,400 words with references) fits inside 40 double-spaced pages only if cut to about 9,000 words all-in. Map below.] + +| § | Section (OT section it condenses) | Words | Change for EURAM | +|---|---|---|---| +| 1 | Introduction (OT Intro) | 900 | Lead with OB/HRM stakes: appraisal and hiring through AI. | +| 2 | The triadic view of platforms (OT §2) | 1,200 | Keep the seller and hotel cases; cut the ride-hail sentence. | +| 3 | What stands between (OT §3) | 1,400 | Keep Simmel's mediator–arbitrator contrast and the oracy → literacy → algorithmacy line. | +| 4 | Algorithmacy (OT §4) | 2,000 | Three practices and two failure modes. Add one sentence each on training implications. | +| 5 | Discussion (OT §5) | 1,300 | Lead with deskilling plus relational demand, which suits ST09_03's job-quality concern. | +| 6 | Conclusion and agenda (OT §6) | 500 | Keep the agenda on asymmetric triads and GenAI in professional work. | +| | References | ~1,700 | About 70 entries | +| | **Total** | **~9,000** | | diff --git a/submissions/euram2027_algorithmacy/README.md b/submissions/euram2027_algorithmacy/README.md new file mode 100644 index 00000000..f4fce010 --- /dev/null +++ b/submissions/euram2027_algorithmacy/README.md @@ -0,0 +1,18 @@ +# EURAM 2027: Algorithmacy for an OB/HRM audience + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. The track choice is the author's.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and condensation outline drafted. **A full paper (10–40 pages) is due 12 Jan 2027 at 14:00 Brussels time.** Nothing has been submitted. + +**The call:** EURAM 2027, Amsterdam, 22–25 June 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Submission platform opens | late November 2026 | +| Full paper and abstract (≤400 words) | **12 Jan 2027, 14:00 Brussels** | +| Notification | 11 Mar 2027 | +| Registration of presenting author | 22 Apr 2027 | + +**Angle:** the OT paper itself, condensed for ST09_03. + +**Relation to other arms:** a short form of [`../algorithmacy_ot/`](../algorithmacy_ot/). FLAGS 2 covers the concurrency question. diff --git a/submissions/euram2027_algorithmacy/cfp/README.md b/submissions/euram2027_algorithmacy/cfp/README.md new file mode 100644 index 00000000..cf726308 --- /dev/null +++ b/submissions/euram2027_algorithmacy/cfp/README.md @@ -0,0 +1,34 @@ +# CFP brief: EURAM 2027 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://conferences.euram.academy/2027conference/call-for-papers/ +- https://conferences.euram.academy/2027conference/authors-guidelines-for-full-papers/ +- https://conferences.euram.academy/2027conference/sig-09-ob/ + +**Event:** EURAM 2027, "Rising Above: Embracing Resilience, Reimagining History, and Rethinking Management for a Sustainable Future". Vrije Universiteit Amsterdam, 22–25 June 2027. + +## Dates (verbatim) +- "Paper submission opens: Late November 2026" +- "Deadline for Paper Submission: 12 January 2027, 2.00 pm Brussels time" +- "Notification of Acceptance: 11 March 2027" +- "Authors and Panelists Registration Deadline: 22 April 2027" + +## Format (verbatim) +- "The maximum length of the paper is 40 pages (including ALL tables, appendices and references). The minimum length of the paper is 10 pages." +- "Use Times New Roman 12-point font, double-spaced, and 1-inch (2.5 cm) margins all around." +- "The abstract should not exceed 400 words and should be included both in the PDF and the submission form." +- 3 to 6 keywords; one anonymised PDF. +- "If you submit a paper, you agree to review at least two papers for EURAM 2027 in the track you have submitted your paper to." + +## Originality and limits (verbatim) +- "Submitted papers must NOT have been published or accepted for publication before the EURAM annual Conference (neither in print nor online, in any form – online preprint, working paper, etc.)" +- "Each paper can only be submitted to ONE track." +- "ONE PRESENTING AUTHOR = ONE PAPER" +- The guidelines say nothing about papers under review at a journal. + +## Tracks +- **ST09_03, "AI, Big Data, Algorithmic Management and Emerging Tech in Managing People"** (Yumei Yang): "Key concerns include impacts on health, job quality, autonomy, and fairness in algorithmic decision-making. Ethical issues surrounding transparency and accountability further complicate adoption." +- **T09_04, "Distance, Voice and Good Work in Digital Collaboration"** (Sylvia Rohlfer): "We welcome conceptual, empirical, and methodological contributions on cognitive, socio-relational, operational, and engagement-related distances in remote working and collaboration; competence development at different levels". +- **T09_07, "Human, Artificial, and Organizational Cognition"** (Davide Secchi): "environments where a mix of human and artificial agents can interact but maintain separate individualities". diff --git a/submissions/facct2027_triad_accountability/ABSTRACT.md b/submissions/facct2027_triad_accountability/ABSTRACT.md new file mode 100644 index 00000000..4230cde7 --- /dev/null +++ b/submissions/facct2027_triad_accountability/ABSTRACT.md @@ -0,0 +1,24 @@ +# Abstract: FAccT 2027 (registration, non-archival) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Registration closes 27 Oct 2026, 11:59 PM AoE. The title and abstract can be edited until 3 Nov.] + +**Title:** Accountability in the Triad: Who Answers When an Algorithm Judges Two People at Once + +## Abstract +Algorithmic accountability research models a two-place relation: a system on one side and an affected person on the other. Many systems do not sit in that position. A marketplace ranking, a booking platform's review score or a hiring screen stands between two people who depend on each other. It evaluates both of them and binds both to its verdict. We call this arrangement a triad, and we argue that two-place accountability cannot assign responsibility inside it. Following Bovens, accountability is a relation in which an actor explains its conduct to a forum that can judge it. In a triad, each party's outcome depends on the counterpart and on the rule at once. The evaluated party cannot tell how much of a verdict each one produced, and the operator can deflect responsibility onto either party. A forum that hears only the affected person therefore hears half of the evidence. We draw on published studies of seller ratings, hotel rankings and host evaluations on home-sharing platforms to specify what a three-place forum requires. It must hear both parties together. It must also support three competencies that let participants attribute responsibility: interpreting an ambiguous result, specifying intent so the counterpart can read it through the system, and tracking rule changes the operator does not announce. We close with design and audit implications. Audits should test whether a system lets the two parties jointly contest one verdict. They should also test whether its explanations separate the counterpart's contribution from the rule's. + +**Keywords:** algorithmic accountability; platforms; contestability; triads; organizational factors + +**Word count:** 244 (body, by `wc -w`). + +## Citation check +| Claim in the abstract | Source card | Depth | +|---|---|---| +| Accountability as an actor explaining to a forum that can judge | `lima_pdw/literature/cards/bovens2007.md` | full text | +| The evaluated party cannot split a verdict between client and platform | `lima_pdw/literature/cards/rahman2021.md` (p. 976, paraphrase) | full text | +| The operator deflects responsibility | `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) | full text | +| Seller ratings enact a triad | `lima_pdw/literature/cards/curchod2020.md` | full text | +| Hotel rankings reconfigure accountability | `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` | full text | +| Folk theories revised as rules change | `lima_pdw/literature/cards/devito2021.md` | extended | + +FAccT abstracts carry no reference list. The paper cites these in ACM format. diff --git a/submissions/facct2027_triad_accountability/ABSTRACT_REGISTERED.md b/submissions/facct2027_triad_accountability/ABSTRACT_REGISTERED.md new file mode 100644 index 00000000..3815914a --- /dev/null +++ b/submissions/facct2027_triad_accountability/ABSTRACT_REGISTERED.md @@ -0,0 +1,9 @@ +# Abstract as registered on OpenReview, 26 Oct 2026 (author's wording) + +**Title:** Accountability in the Triad: Who Answers When an Algorithm Judges Two People at Once + +**TL;DR:** When an algorithm evaluates two interdependent people at once, two-place accountability fails, so platforms need a triadic forum that hears both parties together and separates each one's influence from the algorithm's. + +Algorithmic accountability research typically models a two-place relationship, with a decision system on one side and an affected individual on the other. However, many socio-technical systems don't fit this dyadic structure. For example, marketplace rankings, ride-hailing matches, and booking platform review scores mediate interactions between two interdependent parties, evaluating both simultaneously and binding each to the outcome. This arrangement is called a triad, an organizational structure in which conventional accountability models struggle to assign responsibility. According to Bovens' conceptual framework, accountability is a relational mechanism in which an actor must explain and justify conduct to a forum authorized to judge it. In a triad, each party's outcome depends on both the counterpart's behavior and the platform's proprietary rules. The evaluative algorithm often obscures how much each party contributes to the outcome, allowing the platform operator to shift responsibility onto either user. As a result, any accountability forum that considers only one party receives incomplete evidence. By synthesizing empirical findings from online seller ratings, hotel ranking systems, and home-sharing host evaluations, this work specifies the requirements for a triadic forum. Such a forum must adjudicate both parties simultaneously and support three essential competencies for contesting responsibility: interpreting ambiguous algorithmic feedback, specifying communicative intent through the system, and tracking unannounced rule changes. The paper concludes with design and auditing implications, emphasizing that audits should assess whether systems enable joint contestation and whether explanations effectively separate counterpart influence from algorithmic intervention. + +**Keywords (as entered):** Algorithmic contestability; Triads; Organizational factors; Algorithmic accountability; Multi-party coordination diff --git a/submissions/facct2027_triad_accountability/ARGUMENT.md b/submissions/facct2027_triad_accountability/ARGUMENT.md new file mode 100644 index 00000000..56920a32 --- /dev/null +++ b/submissions/facct2027_triad_accountability/ARGUMENT.md @@ -0,0 +1,20 @@ +# Argument: Accountability in the triad + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling. Sources are card paths in this repo.] + +**P1.** Accountability is a relation between an actor and a forum. The actor must explain and justify its conduct, and the forum can question it and pass judgment with consequences. +- *Warrant:* this is the standard definition in the accountability literature, and FAccT work on algorithmic accountability uses it. +- *Sources:* `lima_pdw/literature/cards/bovens2007.md` (full text); `lima_pdw/literature/cards/wieringa2020.md` (abstract depth, for the FAccT uptake only). + +**P2.** When an algorithmic intermediary evaluates two parties and binds them both, neither party is a forum for the intermediary, and each party's outcome depends on the intermediary and the counterpart at once. Its rules are also opaque in a way that disclosure does not cure. +- *Warrant:* on marketplaces and booking sites the operator turns one party's reviews into a verdict on the other, and the evaluated party cannot see how much of the verdict came from the platform and how much from the counterpart. The operator can then distribute, deflect and deny accountability. +- *Sources:* + - `lima_pdw/literature/cards/curchod2020.md` (triad among operator, buyers and sellers) + - `lima_pdw/literature/cards/rahman2021.md` (p. 976, the five components of the invisible cage; the evaluated worker cannot tell how much client and platform each contributed) + - `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14: accountability "distributed, deflected, and denied") + - `lima_pdw/literature/cards/burrell2016.md` (three kinds of opacity, the third intrinsic to machine learning at scale; abstract depth) + - `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` (ratings subvert hierarchical accountability in travel) + +**C.** Accountability for an AI-mediated relationship must be specified as a three-place relation. It needs a forum that both parties can call on together. It also needs the parties to have the competence to tell whether an outcome came from the counterpart or from the rule: interpreting results, specifying intent across the filter, and tracking rule changes. Two-place accountability (system to affected person) cannot assign responsibility in a triad, because it hears only one of the two people whose joint work the system judged. +- *Warrant for the step:* if P1 defines accountability by an actor–forum pair, and P2 shows that a triad has two affected parties and no shared forum, then a two-place mechanism leaves out the counterpart whose conduct is half the evidence. +- *Sources for the three competencies:* the OT manuscript (under review; not cited in an anonymized submission, see FLAGS); `lima_pdw/literature/cards/devito2021.md` (folk theories revised as platforms change). diff --git a/submissions/facct2027_triad_accountability/FLAGS.md b/submissions/facct2027_triad_accountability/FLAGS.md new file mode 100644 index 00000000..881539a2 --- /dev/null +++ b/submissions/facct2027_triad_accountability/FLAGS.md @@ -0,0 +1,20 @@ +# Flags: FAccT 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +## Deadline and format +1. **A full paper is needed by 3 Nov 2026 AoE (14 pages).** Registration on 27 Oct only reserves the slot, so registering commits the author to writing about 8,000 words in eight days. +2. **The non-archival choice must be made at registration and cannot be changed.** The OT paper will be under review at *Organization Theory*. FAccT allows conference versions of papers under journal review only as non-archival. This paper is a spinoff, not the OT paper, but choosing non-archival keeps both routes safe. +3. **Authors cannot be changed after registration.** OpenReview needs an account. FAccT recommended creating it by 16 Oct for authors without an institutional affiliation. +4. **At least one author must attend in Porto, 21–24 June 2027.** + +## Evidence +5. **Three sources are at abstract depth only:** Burrell 2016, Wieringa 2020 and Ananny & Crawford 2018. They cannot be quoted or paged until read in full. +6. **Curchod et al. 2020** carries no internal page numbers on its card, so its "triadic relationships" sentence cannot be quoted with a page. +7. **Simmel's arbitrator is chosen by the parties.** On platforms neither party chooses the arbiter. Say so rather than equate the two (claim check S111, S190, S194). +8. **Stark & Vanden Broeck** support the triangle, co-optation and "distributed, deflected, and denied" (p. 14). They do not support "without asking" or "binding outcomes" (claim check). + +## Fit +9. FAccT rules out work "focused on purely hypothetical concerns". The cases must be published empirical studies, never invented vignettes. +10. **Example rule:** no driver or freelancer examples lead the paper (author ruling for the OT paper). Rahman's freelancers appear only as evidence for the unspecified share of a score. +11. **Overlap with the OT paper:** this paper's contribution is the forum argument, which the OT paper does not make. The three competencies are the OT paper's. Here they appear only in §6, as presupposed, and are not claimed as new. diff --git a/submissions/facct2027_triad_accountability/OUTLINE.md b/submissions/facct2027_triad_accountability/OUTLINE.md new file mode 100644 index 00000000..868d5196 --- /dev/null +++ b/submissions/facct2027_triad_accountability/OUTLINE.md @@ -0,0 +1,17 @@ +# Outline: Accountability in the Triad (FAccT 2027, 14 pages) + +[DRAFTED by Claude, 26 Oct 2026. Budget about 8,000 words of body, which is about 13 pages in the ACM single-column template. References and the endmatter page are excluded from the limit.] + +| § | Section | Words | What it does (mechanism) | Cards | +|---|---|---|---|---| +| 1 | Introduction | 900 | States the claim: two-place accountability cannot assign responsibility when a system judges two interdependent people. Opens on one case: a seller demoted after a buyer's review. | `curchod2020`, `starkvandenbroeck2024` | +| 2 | Two-place accountability in FAccT | 1,200 | Shows that the actor–forum relation, explanation, recourse and contestability each assume one system facing one affected person. Names what that model leaves out: the counterpart. | `bovens2007`, `wieringa2020` (abstract depth), `ananny2018` (abstract depth), `kroll2017` | +| 3 | The triad | 1,400 | Defines it: two parties whose outcomes depend on each other, and an intermediary that evaluates both and binds both. Separates a channel (passes messages) from an arbiter (issues verdicts), following Simmel's mediator–arbitrator contrast. Gives three published cases: marketplace sellers, hotel rankings and freelancer scores. | `simmel1902number2` (pp. 167–172), `curchod2020`, `scottorlikowski2012accountability`, `rahman2021` | +| 4 | Why the forum fails | 1,300 | Covers three mechanisms. (a) Confounded evidence: neither party can split a verdict between counterpart and rule. (b) Deflection: the operator assigns blame to either side. (c) No joint standing: each party can complain only alone. | `rahman2021` (p. 976), `starkvandenbroeck2024` (p. 14), `burrell2016` (abstract depth; cite for the three kinds of opacity only) | +| 5 | A three-place forum | 1,400 | Sets out requirements: joint contestation of one verdict; explanations that separate the counterpart's input from the rule's; notice of rule changes; logs both parties can read. | `bovens2007`, `kroll2017`, `ananny2018` | +| 6 | Competencies the forum presupposes | 1,000 | Covers interpreting an ambiguous result, specifying intent across the filter and tracking unannounced change. Each one is tied to the forum requirement it makes usable. | `devito2021`, `bucherschouwaldkirch2021`, `daft1986` | +| 7 | Implications for audit and design | 600 | Gives audit questions for triadic systems, plus a design checklist. | — | +| 8 | Limitations | 200 | The argument is conceptual and the cases are secondary. It does not cover symmetric triads (both parties rated equally). | — | +| | **Total** | **8,000** | | | + +Endmatter (outside the limit): ethics statement, adverse impacts, and the generative AI usage statement. diff --git a/submissions/facct2027_triad_accountability/README.md b/submissions/facct2027_triad_accountability/README.md new file mode 100644 index 00000000..b8acfeec --- /dev/null +++ b/submissions/facct2027_triad_accountability/README.md @@ -0,0 +1,29 @@ +# FAccT 2027: Accountability in the Triad + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. The angle is a proposal for the author to rule on.] + +**Status: abstract REGISTERED 26 Oct 2026** on OpenReview (non-archival; primary focus Power and practice, secondary Normative foundations). The registered abstract is in `ABSTRACT_REGISTERED.md`. **The full 14-page paper is due 3 Nov 2026, 11:59 PM AoE.** + +**The call:** ACM Conference on Fairness, Accountability, and Transparency, Porto, 21–24 June 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Abstract registration (title, abstract, authors, archival choice) | **27 Oct 2026, 11:59 PM AoE** | +| Full paper, 14 pages | 3 Nov 2026, 11:59 PM AoE | +| First-round decisions | 22 Dec 2026 | +| Revision and rebuttal | 28 Jan 2027 | +| Final decisions | 23 Mar 2027 | + +**Angle:** Accountability research assumes one system facing one affected person. When a system judges two interdependent people, a three-place forum is needed. + +**Relation to other arms:** spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/). The forum argument is new here; the three competencies are borrowed. See FLAGS 11. + +| File | What it is | +|---|---| +| `ABSTRACT.md` | Title and 244-word abstract, with its citation check | +| `ARGUMENT.md` | P1, P2, C, with warrants and card paths | +| `OUTLINE.md` | 8-section outline, 8,000 words | +| `FLAGS.md` | Deadline, archival, evidence and fit cautions | +| `cfp/` | Digest; `source/` is gitignored | + +**Guardrails:** Claude does not register or submit. The author's read-aloud is the last gate. Every citation maps to a card. diff --git a/submissions/facct2027_triad_accountability/cfp/README.md b/submissions/facct2027_triad_accountability/cfp/README.md new file mode 100644 index 00000000..b34621ea --- /dev/null +++ b/submissions/facct2027_triad_accountability/cfp/README.md @@ -0,0 +1,34 @@ +# CFP brief: FAccT 2027 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://facctconference.org/2027/cfp.html (page stamp "Updated Tue, 20 Oct 2026") +- https://facctconference.org/2027/authorguide.html ("Last Updated: August 2026") + +## Dates (verbatim) +- "All deadlines are at 11:59 PM Anywhere On Earth on the given date." +- "Abstract submission deadline: October 27, 2026" +- "Paper submission deadline: November 3, 2026" +- "Reviews and first round decisions released to authors …: December 22, 2026" +- "Revision and rebuttal due: January 28, 2027" +- "Final notification of accept/reject decisions: March 23, 2027" +- Conference: "June 21-24 2027", Porto. + +## Format (verbatim) +- "Submissions must be no longer than 14 pages, excluding references (most papers are between 12 and 14 pages)." +- An extra page is allowed for "ethics, adverse impacts, and other statements". +- The ACM single-column template (`acmart`, `manuscript,screen,review,anonymous`). Submissions must be anonymized. +- A "Generative AI usage statement" is required. +- Submission is through OpenReview. "the registered list of authors cannot be modified." +- The abstract field has no stated length limit. The title and abstract can be changed until the paper deadline. + +## Archival choice (verbatim) +- "Non-archival papers will only appear as abstracts in the proceedings". +- The archival or non-archival choice is made "at the time the paper is registered (at the abstract submission deadline)". +- Allowed under the non-archival option: "Conference-length versions of papers that are already under review at a journal, but which have not yet been published in that journal: Authors interested in this option must select the non-archival option". + +## Scope lines nearest this paper +- Topics: "Organizational factors in fairness, accountability, and transparency"; "Resistance, refusal, and contestation of computational systems"; "Labor and economic impacts of computational systems"; "Interpretability/explainability". +- "Research from all fields is welcome, including algorithmic, critical, human-centered, legal, philosophical, sociological, statistical, and theoretical research." +- Out of scope: work "focused on purely hypothetical concerns". diff --git a/submissions/ilpc2027_cooptation_control/ABSTRACT.md b/submissions/ilpc2027_cooptation_control/ABSTRACT.md new file mode 100644 index 00000000..24fb6e99 --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/ABSTRACT.md @@ -0,0 +1,29 @@ +# Abstract: ILPC 2027, general stream ("Technology, control the corporation and autonomy") + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. The call asks for 350–500 words plus keywords. Due 31 Oct 2026; the platform shows 8 Nov.] + +**Title:** Co-optation as Control: The Labour Process Between Two Rated Parties + +## Abstract +Labour process theory locates control in the employment relation. When customers enter the analysis, they enter as an instrument of management: the employer uses customer feedback to discipline workers. Platform work breaks this picture. A marketplace, booking site or service platform employs neither the provider nor the customer. It enrols both in a rating system, turns each side's judgement into a verdict on the other, and binds both to its rules. Stark and colleagues call this co-optation, which differs from command, contract and collaboration. Selznick's co-optation absorbed outsiders into the leadership of an organization. Platform co-optation does the reverse: it projects evaluative authority over people who never chose the platform and have no seat in it. + +This paper asks where the frontier of control runs under co-optation, and what autonomy remains to the people on each side of it. **Theoretical orientation:** labour process theory's contested terrain, read through organizational theories of co-optation and of the third party. **Methodology:** a conceptual analysis built on published ethnographies and case studies of online marketplaces and hotel rating platforms. The paper reads each one for how evaluation passes between the two parties and the operator. + +**Findings:** Three mechanisms recur. First, the operator fuses the counterpart's judgement and its own rule into one score, so a worker cannot tell which one acted. Second, every message to the counterpart is also a message to the evaluator, so the worker addresses two audiences at once. Third, the rule changes without notice, so the worker must infer change from results. Under these conditions, task simplification and relational demand rise together. The simpler the task, the more a worker's standing depends on reading the evaluator, making intent legible through it, and tracking its changes. + +**Contribution:** The frontier of control now runs between the two co-opted parties as well as between worker and employer. Autonomy there takes two forms. One is individual competence in reading and addressing an evaluator. The other is collective voice, which matters because individual exit forfeits ratings that do not transfer between platforms. The paper argues that labour process research should study both forms, and that worker organising on platforms should treat the counterpart as a possible ally, not only a source of discipline. + +**Keywords:** co-optation; algorithmic control; platforms; customer evaluation; voice; conceptual analysis + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Customer feedback as a management instrument | **no card yet**: Fuller & Smith (1991) must be read and carded (FLAGS 3). `lima_pdw/literature/cards/wood2019.md` supports only that client ratings accumulate and cannot be contested | — | +| Co-optation as distinct from command, contract and collaboration | `dxd_foundation/literature/library/stark2020algorithmic.md` (p. 48); `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 1) | full text | +| Selznick's co-optation | `lima_pdw/literature/cards/selznick1949.md` (p. 13) | extended | +| Marketplace evaluation passes between the parties | `lima_pdw/literature/cards/curchod2020.md` | full text | +| Hotel ratings redistribute accountability | `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` | full text | +| Score cannot be split between client and platform | `lima_pdw/literature/cards/rahman2021.md` (p. 976) | full text | +| Ratings do not travel | `hospitality_phygital/library/cards/zervas2021reputation.md` | full text | +| Counterpart as potential ally | `dxd_foundation/literature/library/cruz2023digital.md` (p. 2) | full text | +| Voice | `lima_pdw/literature/cards/hirschman1970.md` (p. 30) | full text | diff --git a/submissions/ilpc2027_cooptation_control/ARGUMENT.md b/submissions/ilpc2027_cooptation_control/ARGUMENT.md new file mode 100644 index 00000000..0c94dc8b --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/ARGUMENT.md @@ -0,0 +1,22 @@ +# Argument: Co-optation as control + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Labour process research locates control in the relation between employer and worker. Where it brings in customers, it treats them as an instrument the employer uses: management by customers. +- *Warrant:* the contested-terrain tradition and its algorithmic extension name control mechanisms that run from the employer to the worker. +- *Sources:* + - `lima_pdw/literature/cards/kellogg2020.md` (full text; six mechanisms) + - Fuller & Smith (1991), management by customers (**no card**; see FLAGS 3); `lima_pdw/literature/cards/wood2019.md` (full text; client ratings accumulate and cannot be contested) + +**P2.** Platforms co-opt two parties at once. They enrol customers and providers in rating and ranking without making either an employee or an agent, then turn each side's evaluation into a verdict on the other. Selznick's co-optation absorbed outsiders into leadership; platform co-optation projects evaluative authority over outsiders. +- *Warrant:* the platform's own operators describe and design this three-sided structure, and marketplace studies document it. +- *Sources:* + - `lima_pdw/literature/cards/selznick1949.md` (p. 13) + - `dxd_foundation/literature/library/stark2020algorithmic.md` (pp. 48, 53) + - `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) + - `lima_pdw/literature/cards/curchod2020.md` + - `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` + +**C.** Under co-optation, the frontier of control runs through the relation between the two co-opted parties, not only through the employment relation. Workers' autonomy then depends on a relational competence: telling the counterpart's judgement from the rule's, making intent legible through the system, and noticing silent rule changes. It also depends on collective voice, because individual exit forfeits accumulated ratings. +- *Warrant for the step:* if control is exercised by binding two parties to each other's evaluations (P2), then the space workers can contest lies between them. That space can be held individually by competence or collectively by voice. +- *Sources:* `lima_pdw/literature/cards/hirschman1970.md` (voice, p. 30); `hospitality_phygital/library/cards/zervas2021reputation.md` (ratings do not travel across sites). diff --git a/submissions/ilpc2027_cooptation_control/FLAGS.md b/submissions/ilpc2027_cooptation_control/FLAGS.md new file mode 100644 index 00000000..f571932a --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: ILPC 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **Deadline conflict.** The call page says 31 Oct 2026 and the platform says 8 Nov 2026, 23:00. Plan on 31 Oct. +2. **Conceptual papers must show "how they address and advance relevant debates."** The abstract names the debate (contested terrain and the customer) and the advance (the frontier between the two parties). +3. **Fuller & Smith (1991) has no card, and the Wood et al. 2019 card does not mention it.** The abstract's first premise (customers as the employer's instrument) rests on this uncarded source. Read and card it before submitting, or soften the sentence. +4. **Braverman, Burawoy, Edwards and Lopez (2010) have no cards.** An ILPC audience expects the first three. Card them before the full paper. +5. **H5 REFRAME** (`coordinative_sovereignty/research/research_database.md`, line 23): co-optation is not a fourth coordination logic. The abstract cites Stark for the mechanism only. It does not claim a new form of organization. +6. **Do not write that Stark & Vanden Broeck say "without asking".** Their word is "folded into" (claim check). The abstract says "never chose the platform", which is the author's claim. +7. **Example rule:** cases are marketplaces and hotels, not drivers or couriers. Rahman's freelancers appear only as evidence for the unsplittable score. +8. **The "possible ally" point** comes from Cruz & Gameiro's mini-review (p. 2), which has no data of its own. diff --git a/submissions/ilpc2027_cooptation_control/OUTLINE.md b/submissions/ilpc2027_cooptation_control/OUTLINE.md new file mode 100644 index 00000000..0917ba58 --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Co-optation as Control (ILPC 2027 conference paper) + +[DRAFTED by Claude, 26 Oct 2026. ILPC asks only for an abstract now. This outline sizes a conference paper at about 7,000 words for the presentation and any later journal version.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 700 | States the claim: the frontier of control runs between two co-opted parties. | `starkvandenbroeck2024`, `curchod2020` | +| 2 | Control and the customer in labour process theory | 1,200 | Covers the contested terrain and management by customers, where the customer is the employer's instrument. | `kellogg2020`, Fuller & Smith 1991 (needs a card), `wood2019` | +| 3 | From absorbing outsiders to rating them | 1,100 | Contrasts Selznick's co-optation with platform co-optation: the direction of authority and the absence of a seat. | `selznick1949`, `stark2020algorithmic`, `starkvandenbroeck2024` | +| 4 | Three mechanisms | 1,600 | Fused verdict, double address and silent rule change, each shown in a marketplace or hotel case. | `curchod2020`, `scottorlikowski2012accountability`, `rahman2021` (theory only), `jhaver2018algorithmic` (hosts' "double negotiation") | +| 5 | Deskilled tasks, relational demand | 900 | Simpler tasks raise the relational load. Contrasts this with the deskilling thesis. | `kellogg2020`; Braverman (needs a card) | +| 6 | Two forms of autonomy | 1,000 | Individual competence and collective voice. Exit is costly because ratings do not travel. The counterpart is a possible ally. | `hirschman1970`, `zervas2021reputation`, `cruz2023digital`, `johnston2018organizing` | +| 7 | Conclusion and agenda | 500 | Sets questions for labour process research on two-sided rating. | — | +| | **Total** | **7,000** | | | diff --git a/submissions/ilpc2027_cooptation_control/README.md b/submissions/ilpc2027_cooptation_control/README.md new file mode 100644 index 00000000..8e002e60 --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/README.md @@ -0,0 +1,27 @@ +# ILPC 2027: Co-optation as Control + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. The angle is a proposal for the author to rule on.] + +**Status: abstract SUBMITTED 26 Oct 2026** on eventos.uca.es (General stream). Decisions are expected in early December 2026. + +**The call:** 45th International Labour Process Conference, Cádiz (Jerez), 31 Mar – 2 Apr 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Abstract, 350–500 words, plus keywords | **31 Oct 2026** (platform: 8 Nov 2026, 23:00) | +| Decisions | early December 2026 | +| Conference | 31 Mar – 2 Apr 2027 | + +**Angle:** platforms co-opt two parties and bind each to the other's evaluation, so the frontier of control runs between them. + +**Relation to other arms:** spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/) (its deskilling paragraph). It shares sources with [`../sase2027_platform_cooptation/`](../sase2027_platform_cooptation/), but the SASE paper argues governance and this one argues control. + +| File | What it is | +|---|---| +| `ABSTRACT.md` | Title, abstract, keywords and citation check | +| `ARGUMENT.md` | P1, P2, C | +| `OUTLINE.md` | 7-section outline, about 7,000 words | +| `FLAGS.md` | Cautions | +| `cfp/` | Digest; `source/` is gitignored | + +**Guardrails:** Claude does not submit. The author's read-aloud is the last gate. diff --git a/submissions/ilpc2027_cooptation_control/cfp/README.md b/submissions/ilpc2027_cooptation_control/cfp/README.md new file mode 100644 index 00000000..41d57fb5 --- /dev/null +++ b/submissions/ilpc2027_cooptation_control/cfp/README.md @@ -0,0 +1,25 @@ +# CFP brief: ILPC 2027 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://www.ilpc.org.uk/ILPC-2027/Call-For-Paper +- https://www.ilpc.org.uk/ILPC-2027/Streams (and streams 1–3) +- https://eventos.uca.es/event_detail/155800/detail/international-labour-process-conference-2027-i-31st-march-n-2nd-april-2027.html (the submission platform) + +**Event:** 45th International Labour Process Conference, "Spaces of Work and Resistance". University of Cádiz, Jerez campus, 31 March – 2 April 2027. + +## Dates: the two pages disagree +- Call page: "Abstract submission through the ILPC 2027 platform will open at the start of September 2026 with a deadline of 31st October 2026. Decisions of acceptance will be made by early December 2026." +- Platform: "Nov 8th '26 23:00 Abstracts submission closing". No timezone is given. +- **Treat 31 Oct 2026 as binding** until the organizers say otherwise. + +## Format (verbatim) +- "Abstracts are externally evaluated and should be between 350 and 500 words. Key words should be given that indicate the focus of research and the methods used. The abstract should contain clear information about theoretical orientation, findings, methodology, and what contribution is being made to knowledge." +- "Abstracts of papers that are concerned solely with theoretical or conceptual matters will need to provide clear information on how they address and advance relevant debates." +- Submit to the general stream or a special stream through the platform. + +## Stream lines nearest this paper +- General stream: "Technology, control the corporation and autonomy". +- Conference theme: "how digital infrastructure reconfigures the spatial dimension of labour control". +- Special stream 3: "New technologies of labour monitoring and surveillance"; "Use of intermediaries"; "The session invites theoretical analyses of the manufacturing of migrant worker consent". diff --git a/submissions/io_algorithmic_assemblages/FLAGS.md b/submissions/io_algorithmic_assemblages/FLAGS.md new file mode 100644 index 00000000..3856da69 --- /dev/null +++ b/submissions/io_algorithmic_assemblages/FLAGS.md @@ -0,0 +1,7 @@ +# Flags: *Information and Organization*, Algorithmic Assemblages + +[DRAFTED by Claude, 26 Oct 2026.] + +1. **The call is probably closed.** The aggregator gives 15 May 2026 for full papers. +2. **The publisher page is blocked to automated reads.** Only a browser read can settle the date. +3. **No drafting was done,** per the stopping rule for unverifiable calls. diff --git a/submissions/io_algorithmic_assemblages/README.md b/submissions/io_algorithmic_assemblages/README.md new file mode 100644 index 00000000..af05da99 --- /dev/null +++ b/submissions/io_algorithmic_assemblages/README.md @@ -0,0 +1,12 @@ +# *Information and Organization*: Algorithmic Assemblages (stub; call apparently closed) + +[DRAFTED by Claude, 26 Oct 2026. A stub, not a project.] + +**Status:** the aggregator's timeline puts the full-paper deadline at **15 May 2026, which has passed**. First-round decisions are due 31 Oct 2026. A 31 Jan 2027 date appeared only in an ambiguous search listing. No abstract, argument or outline was written, because the call cannot be confirmed open. + +**To reopen:** read the ScienceDirect page (special-issue 325053) in a browser. If a new deadline exists, the interpretive, platform-level angle would be "algorithmic assemblages as triads": a field-level account of how platform assemblages bind two outside parties. That angle must stay distinct from [`../orgsci_agency_triad/`](../orgsci_agency_triad/) and [`../io_hype_ai_literacy/`](../io_hype_ai_literacy/). + +| File | What it is | +|---|---| +| `cfp/README.md` | What the aggregator says | +| `FLAGS.md` | Why this is a stub | diff --git a/submissions/io_algorithmic_assemblages/cfp/README.md b/submissions/io_algorithmic_assemblages/cfp/README.md new file mode 100644 index 00000000..e664122d --- /dev/null +++ b/submissions/io_algorithmic_assemblages/cfp/README.md @@ -0,0 +1,20 @@ +# CFP brief: *Information and Organization*, "Algorithmic Assemblages—Fields, Ecosystems, and Platforms: An Interpretive Approach" + +[DRAFTED by Claude, 26 Oct 2026. The publisher page (sciencedirect.com/special-issue/325053) returned HTTP 403. Everything here comes from an aggregator, so verify it in a browser.] + +**Read on 26 Oct 2026:** https://callsforpapers.org/call/io-algorithmic-assemblagesfields-ecosystems-and-platforms-an-interpretive-approach (aggregator). The servicesetu.org listing returned 404. + +**Guest editors (per the aggregator):** Timothy Hannigan, P. Devereaux Jennings, Shaila Miranda, Samer Faraj. Some listings also name Mayur Joshi. + +## Timeline (verbatim, aggregator) +- "August 19, 2025: Call for papers published" +- "May 15, 2026: Submission deadline for full papers" +- "October 31, 2026: First round decision letters sent out" +- "February 28, 2027: First round paper resubmissions" + +A ScienceDirect special-issues listing seen in search results on 26 Oct showed 31 January 2027 near this title. The page layout did not tie that date clearly to this call. + +## Topics (verbatim, aggregator) +- "Interpretive approaches to algorithmic assemblages" +- "Relational approaches from organization theory" +- "The dynamics of algorithmic technologies in organizational ecosystems" diff --git a/submissions/io_hype_ai_literacy/ABSTRACT.md b/submissions/io_hype_ai_literacy/ABSTRACT.md new file mode 100644 index 00000000..cd74f030 --- /dev/null +++ b/submissions/io_hype_ai_literacy/ABSTRACT.md @@ -0,0 +1,18 @@ +# Abstract: *Information and Organization*, Hype Studies + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Due 31 Jan 2027 (aggregator date; verify).] + +**Title:** Selling Literacy for a Triadic Problem: AI Literacy as an Organized Remedy + +## Abstract +Hype attaches to remedies as well as to technologies. Alongside the promises made for generative AI, a second set of promises has grown up around "AI literacy": the claim that people who understand how AI works will be safe, employable and in control. Universities, vendors, consultancies, certification bodies and regulators now produce and mandate AI-literacy programmes. This paper treats that remedy as an object of hype studies. Historians of literacy have shown that print-literacy campaigns promised individual advancement that the social order did not deliver. Recent critics argue that AI-literacy initiatives use literacy uncritically as a metaphor and repeat that promise. We add an organizational argument about the problems the remedy is sold for. Many of them are triadic. A booking platform ranks a hotel on its guests' reviews, a marketplace scores a seller on buyers' ratings, and an AI tool can draft a manager's appraisal of an employee. In each case a system evaluates two people and binds both to its verdict. Knowing how the model works does not tell either party whose input moved the verdict, how a message will be scored, or when the rule changed. A reading remedy for a triadic problem is attractive to its sponsors because it places the deficit in the individual and leaves the arbiter unexamined. We propose a research agenda on the AI-literacy remedy market with four questions. Who produces and certifies the remedy? How do mandates turn it into a compliance obligation? What expectations does it carry? What does it move out of view? + +**Keywords:** hype; AI literacy; literacy myth; algorithmic management; expectations + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Literacy campaigns' promise of advancement | `berthon_ethics_course/literature/cards/graff1991literacy.md` (pp. 2, 321) | partial text | +| AI literacy uses literacy "uncritically as a metaphor" | `berthon_ethics_course/literature/cards/kamperman2026ai.md` (p. 35) | full text | +| AI-literacy frameworks define knowledge of AI | `lima_pdw/literature/cards/long2020.md` | full text | +| Hotel rankings; seller scores; appraisal (GenAI conjecture) | `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3); `lima_pdw/literature/cards/curchod2020.md`; `lima_pdw/literature/cards/hindsvonkrogh2024.md` (p. 4) | full text | diff --git a/submissions/io_hype_ai_literacy/ARGUMENT.md b/submissions/io_hype_ai_literacy/ARGUMENT.md new file mode 100644 index 00000000..585484d1 --- /dev/null +++ b/submissions/io_hype_ai_literacy/ARGUMENT.md @@ -0,0 +1,11 @@ +# Argument: AI literacy as a hyped remedy + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** "AI literacy" is promoted by vendors, universities, policy bodies and training providers as the remedy for the risks of AI at work. It uses literacy as a metaphor and repeats the print-literacy campaigns' promise that individual competence brings security. +- *Sources:* `berthon_ethics_course/literature/cards/kamperman2026ai.md` (pp. 35–36); `berthon_ethics_course/literature/cards/graff1991literacy.md` (pp. 2, 321). + +**P2.** Many of the problems the remedy is sold for are triadic. A system evaluates two people and binds both, and the operator's rule is opaque and changes. A competence defined as knowledge of how AI works (the literacy frameworks) does not reach that relation. +- *Sources:* `lima_pdw/literature/cards/long2020.md`; `dxd_foundation/literature/library/oeldorfhirsch2023what.md` (p. 693: literacy "should not be solely the responsibility of the social media user"); `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14). + +**C.** The hype around AI literacy is organizationally useful because it places the deficit in the individual and leaves the arbiter unexamined. Hype studies should treat the AI-literacy remedy market as an object: who produces the remedy, how it is certified and mandated, and what it moves out of view. diff --git a/submissions/io_hype_ai_literacy/FLAGS.md b/submissions/io_hype_ai_literacy/FLAGS.md new file mode 100644 index 00000000..a452ef7b --- /dev/null +++ b/submissions/io_hype_ai_literacy/FLAGS.md @@ -0,0 +1,11 @@ +# Flags: *I&O* Hype Studies + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The call has been seen only on an aggregator.** Verify the call, the date and the PDW on ScienceDirect in a browser. +2. **A tension with the OT paper.** The OT paper's own image ("handing books to people who cannot read") suggests a literacy-style remedy. This paper must address that directly, for example by saying the OT image is about the missing competence, not about a reading curriculum. It must not quietly contradict the OT paper. +3. **The hype literature is uncarded.** There are no cards for Borup et al. (2006), van Lente, or any hype-studies source. Van Lente is a guest editor. Card the expectations literature before writing §2. +4. **The four producers named** (universities, vendors, consultancies, certification bodies and regulators) are the author's description. No card documents the AI-literacy market yet, so the paper needs sources such as policy documents. +5. **Kamperman & Graff** describe UNESCO and OECD documents without checking them. Their remedy is person-side. +6. **Ng et al. (2021)** is abstract depth only. +7. **Overlap:** this arm critiques the remedy market. It must not propose a measure, which belongs to `../perspsych_measuring_algorithmacy/`, or design properties, which belong to `../isj_future_of_work_design/`. diff --git a/submissions/io_hype_ai_literacy/OUTLINE.md b/submissions/io_hype_ai_literacy/OUTLINE.md new file mode 100644 index 00000000..0920b7ed --- /dev/null +++ b/submissions/io_hype_ai_literacy/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Selling Literacy for a Triadic Problem (*I&O*; length to confirm, planned at about 9,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction: hype attaches to remedies | 900 | Covers the AI-literacy boom and the remedy as an object. | `kamperman2026ai` | +| 2 | Hype and expectations | 1,400 | The sociology of expectations and hype as patterned and consequential. | **needs cards:** Borup et al. 2006; van Lente (FLAGS 3) | +| 3 | The literacy myth, then and now | 1,400 | Print campaigns and AI-literacy campaigns compared. | `graff1991literacy`, `kamperman2026ai`, `goody1987interface` (partial) | +| 4 | What the remedy is sold for | 1,600 | Triadic problems: the system as arbiter between two people. | `starkvandenbroeck2024`, `curchod2020`, `orlikowskiscott2023`, `hindsvonkrogh2024` | +| 5 | Why a reading remedy is attractive | 1,400 | It locates the deficit in the individual, leaves the arbiter unexamined, and suits compliance. | `oeldorfhirsch2023what` (p. 693), `long2020`, `ng2021` (abstract) | +| 6 | Research agenda | 1,600 | Four questions, with methods (document analysis of mandates, an ethnography of certification). | — | +| 7 | Conclusion | 700 | — | — | +| | **Total** | **9,000** | | | diff --git a/submissions/io_hype_ai_literacy/README.md b/submissions/io_hype_ai_literacy/README.md new file mode 100644 index 00000000..f2ef1154 --- /dev/null +++ b/submissions/io_hype_ai_literacy/README.md @@ -0,0 +1,9 @@ +# *Information and Organization*, Hype Studies: Selling Literacy for a Triadic Problem + +[DRAFTED by Claude, 26 Oct 2026. The call has been seen only on an aggregator.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **Due 31 Jan 2027** (aggregator date; verify). There is an online PDW in late 2026 (timing TBA). Nothing has been submitted. + +**Angle:** AI literacy as a hyped remedy that places the deficit in the individual and leaves the arbiter unexamined. + +**Relation to other arms:** draws on [`../berthon_ethics_course/`](../berthon_ethics_course/) (literacy history). It is distinct from [`../algorlit2027_literacy_to_algorithmacy/`](../algorlit2027_literacy_to_algorithmacy/), which addresses literacy researchers, not the remedy market. diff --git a/submissions/io_hype_ai_literacy/cfp/README.md b/submissions/io_hype_ai_literacy/cfp/README.md new file mode 100644 index 00000000..947cb119 --- /dev/null +++ b/submissions/io_hype_ai_literacy/cfp/README.md @@ -0,0 +1,18 @@ +# CFP brief: *Information and Organization*, "Hype Studies: A Research Agenda for Organizing in Digital Futures" + +[DRAFTED by Claude, 26 Oct 2026. **The publisher page (ScienceDirect special-issue 331903) returned HTTP 403.** Everything here comes from an aggregator (servicesetu.org/call-for-paper/journals/iao, posted 9 Apr 2026). Verify in a browser.] + +**Guest editors:** Neil Pollock (University of Edinburgh), Danielle Logue (UNSW Sydney), Harro van Lente (Maastricht University), Marian Gatzweiler (University of Edinburgh). + +## Dates (aggregator, verbatim) +- "Submission deadline: 31 January 2027" +- "An online Paper Development Workshop will be held (late 2026, timing TBA) for feedback prior to formal submission (attendance encouraged but not mandatory)." + +## Format +- Elsevier Editorial Manager, article type "VSI: Hype Studies". Double-anonymized review. +- Length: not stated; see the journal's guide for authors. + +## Lines nearest this paper (verbatim, aggregator) +- "Hype, decision‑making and organisational practice: How do organisations adopt, invest in, and govern technologies amid competing and exaggerated claims?" +- "Intermediaries and the infrastructure of hype: How do analysts, consultants, investors, and media actors produce, evaluate, and professionalise hype?" +- "Contestation, politics and responsibility: How do actors challenge, reframe, or manage "hype" and its risks for equity, governance, and policy?" diff --git a/submissions/isj_future_of_work_design/ABSTRACT.md b/submissions/isj_future_of_work_design/ABSTRACT.md new file mode 100644 index 00000000..72a0cd52 --- /dev/null +++ b/submissions/isj_future_of_work_design/ABSTRACT.md @@ -0,0 +1,19 @@ +# Abstract: *ISJ*, "AI and the Future of Work" + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Due 31 Jan 2027 (indexed date; verify).] + +**Title:** Designing the Third Party: Provenance, Dual Address and Change Notice in Systems That Judge Two People + +## Abstract +AI at work increasingly stands between two people. A booking platform ranks a hotel on its guests' reviews, a marketplace scores a seller on buyers' feedback, and an AI tool can draft the appraisal a manager gives an employee. In each case the system evaluates a relation and binds both parties to its verdict. IS design research has answered algorithmic opacity with explanation, transparency and contestability, and each answer models a single decision subject facing a system. This paper argues that systems which judge two people need different design properties. The parties' problem is not that the model is hidden; machine-learning opacity is intrinsic. Their problem is that they cannot tell whose input moved a verdict, cannot tell how a message to the other party will be scored, and cannot tell when the rule has changed. We derive three design properties from these problems. Provenance cues show which party's input moved a verdict. Dual address shows a sender what the evaluator will register and what the counterpart will see. Change notice logs rule updates and announces them to both parties. We state each property as a testable design proposition. We ground the propositions in published studies of review filters and contestation, and specify outcome measures for each: correct attribution of responsibility, shared interpretation of messages, and detection of rule change. The contribution is a design theory for the human–machine–institution interface in which the machine is a third party, not a tool or a channel. + +**Keywords:** design theory; algorithmic management; contestability; accountability; future of work + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Opacity intrinsic to ML | `lima_pdw/literature/cards/burrell2016.md` | abstract | +| Disclosure is not traceability | `dxd_foundation/literature/library/kroll2021outlining.md` (p. 4) | full text | +| Contestability as a system feature | `dxd_foundation/literature/library/alfrink2023contestable.md` (p. 614) | full text | +| Users write for the filter once told of it | `dxd_foundation/literature/library/eslami2019user.md` (pp. 11–12) | full text | +| Hotel rankings; seller scores | `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3); `lima_pdw/literature/cards/curchod2020.md` | full text | diff --git a/submissions/isj_future_of_work_design/ARGUMENT.md b/submissions/isj_future_of_work_design/ARGUMENT.md new file mode 100644 index 00000000..2a4a5ce7 --- /dev/null +++ b/submissions/isj_future_of_work_design/ARGUMENT.md @@ -0,0 +1,17 @@ +# Argument: design for the third party + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** When an IS artifact evaluates two people and binds both, each person needs to interpret its verdicts, address the other person through it, and track its changes. Transparency about the model does not supply any of these, because the opacity is intrinsic and the parties need to know whose input moved the verdict, not how the model works. +- *Sources:* + - `lima_pdw/literature/cards/burrell2016.md` (abstract depth) + - `dxd_foundation/literature/library/kroll2021outlining.md` (p. 4: disclosure is not traceability) + - `lima_pdw/literature/cards/rahman2021.md` (p. 976: the parties' shares are unspecified) + +**P2.** Design research offers contestability and traceability as system features, but it models one decision subject. Users told that an evaluating filter exists start writing for it as an audience. +- *Sources:* + - `dxd_foundation/literature/library/alfrink2023contestable.md` (p. 614) + - `dxd_foundation/literature/library/lyons2021conceptualising.md` (p. 15: contestation imagined as post-hoc appeal) + - `dxd_foundation/literature/library/eslami2019user.md` (pp. 11–12) + +**C.** IS artifacts that stand between two people need three design properties keyed to the three practices. The first is provenance cues, which show which party's input moved a verdict. The second is a dual address, which shows each sender what the evaluator will read and what the counterpart will read. The third is change notice, which logs and announces rule updates to both parties. Each property can be stated as a testable design proposition without promising transparency of the model. diff --git a/submissions/isj_future_of_work_design/FLAGS.md b/submissions/isj_future_of_work_design/FLAGS.md new file mode 100644 index 00000000..34f96fc2 --- /dev/null +++ b/submissions/isj_future_of_work_design/FLAGS.md @@ -0,0 +1,10 @@ +# Flags: *ISJ*, "AI and the Future of Work" + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The call page was not read.** The date (31 Jan 2027) and the scope come from search-indexed text. Read it in a browser before writing. +2. **Overlap with the OT paper:** the OT paper claims to shift focus "from technical transparency to practical, situated competence". This paper must not restate that as its contribution. Its contribution is the three design properties and their propositions. Avoid an "explainability" framing. +3. **Overlap with `../sij_service_triad_ethics/`:** that paper keeps to governance and organizational mechanisms in one sector. This one owns artifact design properties. +4. **Lab work:** the Spaces in the Stream draft (`../phd1750_spaces_in_stream/`) holds a decision-space design theory that is under development and unpublished. If its ideas are used, cite it as the author's own working paper. Anonymized review then needs care. +5. **Abstract-depth sources:** Burrell 2016, Ananny & Crawford 2018 and DeVito et al. 2017. +6. **Eslami et al. (2019)** report self-reported intentions about writing for the algorithm, not behaviour. diff --git a/submissions/isj_future_of_work_design/OUTLINE.md b/submissions/isj_future_of_work_design/OUTLINE.md new file mode 100644 index 00000000..ffc88ff8 --- /dev/null +++ b/submissions/isj_future_of_work_design/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Designing the Third Party (*ISJ*; length to confirm, planned at about 9,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 900 | Three cases; design answers that assume one subject. | `orlikowskiscott2023`, `curchod2020`, `hindsvonkrogh2024` | +| 2 | Design answers to opacity | 1,400 | Explanation, transparency, traceability and contestability, and the single subject each assumes. | `kroll2021outlining`, `alfrink2023contestable`, `lyons2021conceptualising`, `ananny2018` (abstract) | +| 3 | The parties' three problems | 1,400 | Whose input moved it; how a message will be scored; when the rule changed. | `rahman2021` (theory only), `eslami2019user`, `devito2017` (abstract), `devito2021` | +| 4 | Three design properties | 2,000 | Provenance cues, dual address and change notice, each with a mock-up and its kernel theory. | `eslami2019user`, `kroll2021outlining` | +| 5 | Design propositions and evaluation | 1,500 | Six propositions and outcome measures. Proposes an experimental evaluation (a vignette or simulated marketplace). | — | +| 6 | Discussion | 1,200 | Limits of design under an interested operator. Regulation hooks. | `starkvandenbroeck2024` | +| 7 | Conclusion | 600 | — | — | +| | **Total** | **9,000** | | | diff --git a/submissions/isj_future_of_work_design/README.md b/submissions/isj_future_of_work_design/README.md new file mode 100644 index 00000000..9a84bd06 --- /dev/null +++ b/submissions/isj_future_of_work_design/README.md @@ -0,0 +1,11 @@ +# *ISJ*, "AI and the Future of Work": Designing the Third Party + +[DRAFTED by Claude, 26 Oct 2026. The call page was blocked; facts come from indexed text.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **Due 31 Jan 2027** (indexed date; verify). Nothing has been submitted. + +**The call:** *Information Systems Journal* special issue. Digest: [`cfp/README.md`](cfp/README.md). + +**Angle:** three design properties (provenance cues, dual address, change notice) for systems that judge two people. + +**Relation to other arms:** draws on [`../phd1750_spaces_in_stream/`](../phd1750_spaces_in_stream/) (design theory) and [`../algorithmacy_ot/`](../algorithmacy_ot/) (the three practices). diff --git a/submissions/isj_future_of_work_design/cfp/README.md b/submissions/isj_future_of_work_design/cfp/README.md new file mode 100644 index 00000000..9e907783 --- /dev/null +++ b/submissions/isj_future_of_work_design/cfp/README.md @@ -0,0 +1,13 @@ +# CFP brief: *Information Systems Journal*, "AI and the Future of Work" + +[DRAFTED by Claude, 26 Oct 2026. **The Wiley page (si-2026-000678) returned HTTP 403.** The facts below come from search-engine text of Wiley's own domain, not from a page read. Verify everything in a browser.] + +**URL:** https://onlinelibrary.wiley.com/page/journal/13652575/homepage/call-for-papers/si-2026-000678 + +## What the indexed text says +- Submission deadline: 31 January 2027 (from the journal's call-for-papers listing). +- Invites theoretical and empirical work on how AI reshapes task distribution and delegation, and on workers' voice and agency. +- Focus on "the human-machine-institution interface", organisational governance, power, worker identity and voice, care and invisible labour. + +## Unknown +Guest editors, word limit, article types, any PDW, and the exact wording of the call. diff --git a/submissions/jitcar_post_ai_enterprise/ABSTRACT.md b/submissions/jitcar_post_ai_enterprise/ABSTRACT.md new file mode 100644 index 00000000..0a3a8d31 --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/ABSTRACT.md @@ -0,0 +1,18 @@ +# Abstract: *JITCAR*, "The Post AI Enterprise" + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. **A full paper is due 15 Nov 2026**; the word limit is unknown. This is a 250-word abstract.] + +**Title:** The Internal Third Party: Allocating Algorithmacy in the Post-AI Enterprise + +## Abstract +Generative AI now stands between people inside the firm. It can draft a manager's appraisal of an employee, screen candidates for a hiring manager, and summarize one colleague's work for another. In each case the system evaluates a relation between two of the firm's people, and the firm holds two roles at once. As operator, it chooses and configures the system. As principal, it employs one of the parties as its agent. This paper argues that this double position changes what organizations must do about AI-enabled decisions. On platforms, the competence to coordinate through an evaluating system is left to each participant. Inside the firm it cannot be, because the firm controls what each party sees and who can question a verdict. We call this competence algorithmacy. It has three practices: interpreting an ambiguous result, specifying intent so the other party can read it through the system, and tracking model changes. We argue that the post-AI enterprise must allocate each practice to a role. Someone must interpret verdicts for both parties, someone must hold the right to specify intent to the system, and someone must track updates. Drawing on published studies of knowledge brokers who translate algorithmic outputs, and of experts whose authority erodes when AI presents uncertain projections as certain, we propose an allocation framework and four propositions. The framework treats contestability as a role assignment, not only as a system feature. Firms that leave the three practices unallocated will find that internal AI deflects accountability between their people instead of supporting it. + +**Keywords:** generative AI; algorithmacy; accountability; performance appraisal; knowledge brokers + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| GenAI in manager–employee evaluation and hiring; reduced visibility | `lima_pdw/literature/cards/hindsvonkrogh2024.md` (pp. 4, 6) | full text | +| Brokers who translate algorithmic outputs | `lima_pdw/literature/cards/waardenburg2022.md` | abstract | +| Artificial certainty erodes expert authority | `lima_pdw/literature/cards/leonardileavell2026.md` (pp. 516, 537) | full text | +| Operator at the apex | `lima_pdw/literature/cards/starkvandenbroeck2024.md` | full text | diff --git a/submissions/jitcar_post_ai_enterprise/ARGUMENT.md b/submissions/jitcar_post_ai_enterprise/ARGUMENT.md new file mode 100644 index 00000000..6ac7ba32 --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/ARGUMENT.md @@ -0,0 +1,15 @@ +# Argument: the internal third party + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** When a firm installs generative AI in appraisal, hiring or peer collaboration, the system stands between two of the firm's people. The firm is then at the apex as the operator, and it is also one party's principal. The manager is the firm's agent; the employee is the firm's subordinate. +- *Warrant:* the manager–GenAI–employee cases Hinds and von Krogh describe have this shape, and the firm chooses and configures the system. +- *Sources:* `lima_pdw/literature/cards/hindsvonkrogh2024.md` (pp. 4, 6); `lima_pdw/literature/cards/starkvandenbroeck2024.md` (the operator at the apex, p. 14). + +**P2.** In that position the firm cannot treat the competence to coordinate through the system as each worker's private skill. The firm controls what each party can see and who may question the verdict. Studies of AI in organizations show brokers and experts taking on the interpreting work. +- *Sources:* + - `lima_pdw/literature/cards/waardenburg2022.md` (brokers; abstract depth) + - `lima_pdw/literature/cards/leonardileavell2026.md` (pp. 516, 537: artificial certainty erodes expert authority) + - `lima_pdw/literature/cards/lebovitz2022.md` (engagement with opaque AI; abstract depth) + +**C.** In the post-AI enterprise, algorithmacy is a firm-level capability that the firm must build and allocate. The firm decides who interprets verdicts, who can specify intent to the system, who tracks model updates, and which party can contest. Because the firm is both operator and principal, how it allocates these roles determines whether internal AI restores or dissolves accountability between its people. diff --git a/submissions/jitcar_post_ai_enterprise/FLAGS.md b/submissions/jitcar_post_ai_enterprise/FLAGS.md new file mode 100644 index 00000000..aaa40d03 --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: *JITCAR* + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **A full paper is due 15 Nov 2026.** The length and article types are unknown because the author instructions were blocked. A case-oriented journal may want a case. +2. **High overlap risk with the OT paper.** The OT conclusion already names hiring, appraisal and peer collaboration as triadic. This paper's contribution must be the double position (operator plus principal) and the allocation framework. The setting alone is not a contribution. If the OT revision is still open, the author may shorten that closing paragraph. +3. **Hinds & von Krogh is conjecture.** Write "may", never "does". +4. **Waardenburg et al. 2022 and Lebovitz et al. 2022 are abstract depth only.** +5. **Tong et al. 2021 (AI performance feedback) has no card.** Do not cite it until read. +6. **Leonardi & Leavell** define artificial certainty differently on pp. 516 and 517. Quote one, not a splice. The certainty belongs to lay audiences. +7. **Venue quality:** *JITCAR* is a lower-tier journal. The author may prefer to hold this angle for a stronger venue. +8. **Measurement is not this paper's job.** It belongs to `../perspsych_measuring_algorithmacy/`. diff --git a/submissions/jitcar_post_ai_enterprise/OUTLINE.md b/submissions/jitcar_post_ai_enterprise/OUTLINE.md new file mode 100644 index 00000000..3f56a49e --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/OUTLINE.md @@ -0,0 +1,13 @@ +# Outline: The Internal Third Party (*JITCAR*; length to confirm, planned at about 7,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 700 | Gives three internal cases (appraisal, hiring, peer summary) and states the double position. | `hindsvonkrogh2024` | +| 2 | From platform triads to internal triads | 1,200 | On platforms the operator is outside both parties. Inside the firm it is the operator and one party's principal. | `starkvandenbroeck2024`, `curchod2020`, `hindsvonkrogh2024` | +| 3 | Three practices, three roles | 1,500 | Interpreting goes to an interpreter or broker. Specifying intent goes to the holder of the prompt or configuration. Tracking goes to a model-change steward. | `waardenburg2022` (abstract), `leonardileavell2026`, `hancock2020ai` | +| 4 | Contestability as role assignment | 1,200 | Which party can contest, before whom, and with what record. | `alfrink2023contestable` (p. 614), `bovens2007`, `kroll2021outlining` | +| 5 | Allocation framework and propositions | 1,400 | A 3 × 2 table: practice by party. Four propositions on the effects of misallocation. | `lebovitz2022` (abstract), `hohenstein2020` | +| 6 | Implications and limits | 1,000 | For CIOs and HR. The paper is conceptual and the cases are secondary. | — | +| | **Total** | **7,000** | | | diff --git a/submissions/jitcar_post_ai_enterprise/README.md b/submissions/jitcar_post_ai_enterprise/README.md new file mode 100644 index 00000000..ff8ec8d9 --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/README.md @@ -0,0 +1,18 @@ +# *JITCAR*, "The Post AI Enterprise": The Internal Third Party + +[DRAFTED by Claude, 26 Oct 2026, from the call page. The angle is a proposal for the author to rule on.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **A full paper is due 15 Nov 2026.** Nothing has been submitted. + +**The call:** *Journal of Information Technology Case and Application Research*. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Full paper | **15 Nov 2026** | +| First-round decisions | 15 Jan 2027 | +| Revised papers | 15 Mar 2027 | +| Final decisions | 15 May 2027 | + +**Angle:** inside the firm, the employer is both the AI's operator and one party's principal, so algorithmacy must be allocated to roles. + +**Relation to other arms:** spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/) (its research agenda). It is distinct from [`../misqe_delegation_triads/`](../misqe_delegation_triads/), where the AI is the delegator's agent, not the employer's evaluator. diff --git a/submissions/jitcar_post_ai_enterprise/cfp/README.md b/submissions/jitcar_post_ai_enterprise/cfp/README.md new file mode 100644 index 00000000..01d0e653 --- /dev/null +++ b/submissions/jitcar_post_ai_enterprise/cfp/README.md @@ -0,0 +1,23 @@ +# CFP brief: *JITCAR*, "The Post AI Enterprise" + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the call page, not a substitute for it.] + +**Read on 26 Oct 2026:** https://think.taylorandfrancis.com/special_issues/the-post-ai-enterprise/. The instructions for authors (tandfonline, journal code UTCA) returned HTTP 403. + +**Journal:** *Journal of Information Technology Case and Application Research* (Routledge). **Editors:** Deepak Khazanchi, Christian Haas (University of Nebraska at Omaha), Dawn Owens (UT Dallas), Xiaodan Yu (UIBE). + +## Dates (verbatim) +- "November 15, 2026: Deadline for paper submission (online)" +- "January 15, 2027: First-round decisions" +- "March 15, 2027: Deadline to submit revised papers (online)" +- "May 15, 2027: Provisional/Final decisions" + +## Format +- Word limit, article types and template: **not stated** (the instructions page was blocked). +- "We invite original research, case studies, theory-building, applied and empirical research on (but not limited to) the following topics:" + +## Lines nearest this paper (verbatim) +- "We especially welcome studies that examine how organizations assess the quality, fairness, transparency, and consequences of AI-enabled decisions, particularly in settings where algorithmic recommendations interact with human judgment, institutional constraints, and organizational incentives." +- "Accountability frameworks for autonomous and generative AI systems" +- "Explainability, transparency, and contestability of AI-enabled decisions" +- "Managerial decision-making with autonomous AI agents" diff --git a/submissions/jms_organizing_truth_arbiter/ABSTRACT.md b/submissions/jms_organizing_truth_arbiter/ABSTRACT.md new file mode 100644 index 00000000..9a63d7b0 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/ABSTRACT.md @@ -0,0 +1,20 @@ +# Abstract: *JMS*, "Organizing Truth in an Age of Polarization and AI" + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. JMS limit: 200 words, 4–6 keywords in alphabetical order. Due 3 May 2027.] + +**Title:** Truth by Arbitration: How Evaluating Platforms Settle What Happened Between Two Parties + +## Abstract +When a hotel and a guest disagree about a stay, or a seller and a buyer about a sale, whose account becomes the fact? On platforms, the answer is increasingly the system's. Its review, rating and ranking record settles what happened, and it displaces older apparatuses of evaluation such as inspection and accreditation. We theorize this as truth by arbitration. The system's record is produced under conditions neither party can inspect. Its rules are withheld, its information asymmetries are designed, and each party shapes its report in anticipation of the other and of the filter. The parties' own accounts carry no comparable authority. Building on research on evaluative apparatuses and algorithmic control, we argue that the only counter-record available to the evaluated parties is the shared understanding they build between themselves outside the system's record. We specify how counter-records are made, kept and asserted, and when platforms admit them. We contribute an account of organizational truth in which authority over facts rests with an arbiter, not with experts or stakeholders. + +**Keywords:** algorithmic management; arbitration; evaluation; platforms; shared understanding; truth + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Rankings displace apparatuses of evaluation | `lima_pdw/literature/cards/orlikowskiscott2023.md` (pp. 1, 3) | full text | +| Ratings subvert hierarchical accountability (travel) | `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` | full text | +| Withheld norms | `dxd_foundation/literature/library/rahman2019invisible.md` (p. v) | partial | +| Deliberately created information asymmetries | `dxd_foundation/literature/library/kellogg2020algorithms.md` (p. 384) | full text | +| Reports shaped in anticipation of the other party | `dxd_foundation/literature/library/bolton2013engineering.md` (§II.2, discussion paper) | full text (DP version) | +| Reports shaped for the filter | `dxd_foundation/literature/library/eslami2019user.md` (p. 11) | full text | diff --git a/submissions/jms_organizing_truth_arbiter/ARGUMENT.md b/submissions/jms_organizing_truth_arbiter/ARGUMENT.md new file mode 100644 index 00000000..7edf9eb8 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/ARGUMENT.md @@ -0,0 +1,18 @@ +# Argument: the arbiter's record as organizational truth + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** When a platform evaluates the joint conduct of two parties, its record of what happened becomes the account that counts. Ratings and rankings displaced older apparatuses of evaluation, and the parties' own accounts carry no comparable weight. +- *Sources:* + - `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` (ratings subvert hierarchical accountability) + - `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 1: "displacing institutional apparatuses from their positions of primacy and authority"; p. 3) + - `lima_pdw/literature/cards/leonardileavell2026.md` (p. 537: stakeholders "treated the outputs as binding forecasts") + +**P2.** That record is produced under conditions the parties cannot inspect. Information asymmetries are deliberately created, norms are hidden, and the parties shape their own reports in anticipation of each other and of the filter. +- *Sources:* + - `dxd_foundation/literature/library/kellogg2020algorithms.md` (p. 384) + - `dxd_foundation/literature/library/rahman2019invisible.md` (abstract, p. v) + - `dxd_foundation/literature/library/bolton2013engineering.md` (reciprocity in feedback; discussion-paper version) + - `dxd_foundation/literature/library/eslami2019user.md` (p. 11) + +**C.** Platforms organize truth by arbitration. The system settles which account of a joint event is fact. The only counter-record available to the parties is a shared understanding they build between themselves, outside the system's record. Organizational research on truth should therefore study how evaluated parties build, keep and assert such counter-records, and when platforms admit them. diff --git a/submissions/jms_organizing_truth_arbiter/FLAGS.md b/submissions/jms_organizing_truth_arbiter/FLAGS.md new file mode 100644 index 00000000..5144d6b4 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: *JMS* Organizing Truth + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The abstract must be at most 200 words.** Count it at assembly. Keywords are already alphabetical. +2. **Overlap with the OT paper is medium.** The OT paper maps "specifying intent" to "shared understanding". This paper's new element is epistemic arbitration (the record as fact) and counter-records. It must not present the competency mapping as its own. +3. **Overlap with `../orgsci_agency_triad/`:** do not use "standing" here. +4. **Orlikowski & Scott (2014)** (TripAdvisor, *Organization Science*) has no card. The OT claim check found the OT paper's sentences about it overstated ("cleaning schedules", "quietly"). Use the 2012 and 2023 papers, or card 2014 first. +5. **Bolton et al. (2013)** is carded from the 2009 discussion paper. Do not quote its numbers as the journal article's. +6. **Fradkin et al. (2021)** (Airbnb simultaneous reveal) has no card. It would be central to §5, so card it. +7. **Espeland & Sauder and Martin (2019) have no cards.** Martin is cited in the call itself. +8. **Kornberger et al. (2017)** is metadata only. diff --git a/submissions/jms_organizing_truth_arbiter/OUTLINE.md b/submissions/jms_organizing_truth_arbiter/OUTLINE.md new file mode 100644 index 00000000..9bcebeb9 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Truth by Arbitration (*JMS*, 10,000–13,000 words including references) + +[DRAFTED by Claude, 26 Oct 2026. Body budget about 9,800 words, plus about 2,200 words of references.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 1,100 | A disputed stay; whose account becomes fact. | `scottorlikowski2012accountability` | +| 2 | Truth, evaluation and authority in organizations | 1,800 | Expert authority, evaluative apparatuses and their displacement. | `orlikowskiscott2023`, `leonardileavell2026`, `kornberger2017evaluative` (metadata only; read first) | +| 3 | Truth by arbitration | 2,000 | The arbiter's record as fact; three conditions (withheld rules, designed asymmetry, anticipatory reports). | `kellogg2020algorithms`, `rahman2019invisible`, `bolton2013engineering`, `eslami2019user` | +| 4 | Counter-records | 2,000 | Shared understanding built outside the system; how it is made, kept and asserted. | `jhaver2018algorithmic`, `lee2015` (peer forums), `zervas2021reputation` | +| 5 | When platforms admit counter-records | 1,400 | Dispute processes, simultaneous reveal and appeal. Propositions. | `alfrink2023contestable` | +| 6 | Discussion: polarization and the arbiter | 1,000 | How arbitration interacts with polarized stakeholders. | — | +| 7 | Conclusion | 500 | — | — | +| | **Body** | **9,800** | | | diff --git a/submissions/jms_organizing_truth_arbiter/README.md b/submissions/jms_organizing_truth_arbiter/README.md new file mode 100644 index 00000000..750f4b76 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/README.md @@ -0,0 +1,9 @@ +# *JMS*, Organizing Truth: Truth by Arbitration + +[DRAFTED by Claude, 26 Oct 2026, from the call page.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **Due 3 May 2027 (10,000–13,000 words).** Nothing has been submitted. + +**Angle:** evaluating platforms settle what happened between two parties. The parties' only counter is a shared understanding kept outside the record. + +**Relation to other arms:** spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/). It is distinct from [`../orgsci_agency_triad/`](../orgsci_agency_triad/) (standing) and from [`../sij_service_triad_ethics/`](../sij_service_triad_ethics/) (accountability mechanisms). diff --git a/submissions/jms_organizing_truth_arbiter/cfp/README.md b/submissions/jms_organizing_truth_arbiter/cfp/README.md new file mode 100644 index 00000000..eba92ea8 --- /dev/null +++ b/submissions/jms_organizing_truth_arbiter/cfp/README.md @@ -0,0 +1,24 @@ +# CFP brief: *Journal of Management Studies*, "Organizing Truth in an Age of Polarization and Artificial Intelligence" + +[DRAFTED by Claude, 26 Oct 2026. A working digest, not a substitute for the editors' text.] + +**Read on 26 Oct 2026:** +- https://journalofmanagementstudies.ac.uk/special-issue-call-for-papers-truth-and-organizations-in-the-age-of-polarization/ +- https://journalofmanagementstudies.ac.uk/publish/author-guidance/manuscript-preparation-guidelines/ + +**Guest editors:** Chidiebere Ogbonnaya (King's College London), Claudia Gabbioneta (York), Hannes Leroy (RSM), Johann Fortwengel (King's College London), Patrick Haack (Lausanne). + +## Dates (verbatim) +- "Submission deadline: 3 May 2027" +- There is a revision workshop in late 2027 for authors with a first R&R. Attending the 2026 JMS conference is "not a precondition for submission." + +## Format (verbatim, JMS guidelines) +- "The word count of 10,000-13,000 includes tables, figures, and references." +- "an informative abstract of no more than 200 words ... Provide a list of 4-6 keywords in alphabetical order". +- No footnotes; double-blind review; an AI-use declaration is required. Submit through ScholarOne (mc.manuscriptcentral.com/jmstudies). + +## Lines nearest this paper (verbatim) +- "perceptions of truth in societal and organizational life are mediated by systems whose operations are often difficult to scrutinize and hard to challenge (Kellogg et al., 2020; Martin, 2019)." +- "AI-driven platforms, for example, curate content in ways that influence which narratives gain prominence and which claims come to appear credible or authoritative" +- "how questions of responsibility are discursively contested in sophisticated blame games" +- "Submissions may employ qualitative, quantitative, mixed-method, computational, interpretive, or conceptual designs". diff --git a/submissions/misqe_delegation_triads/ABSTRACT.md b/submissions/misqe_delegation_triads/ABSTRACT.md new file mode 100644 index 00000000..0a5dd375 --- /dev/null +++ b/submissions/misqe_delegation_triads/ABSTRACT.md @@ -0,0 +1,18 @@ +# Abstract: *MISQE* special issue on delegating work to AI agents (Insight format proposed) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Full paper due 1 Mar 2027. **MISQE requires evidence (FLAGS 1).**] + +**Title:** When Your AI Agent Is Also Your Counterpart's: Managing Delegation That Creates a Third Party + +## Abstract +Managers delegate work to AI agents as if they were hiring an assistant: the manager directs, the agent executes, and oversight catches errors. That picture holds while the agent's work stays with the person who delegated it. It breaks when the output crosses to someone else. A developer's coding agent produces changes a reviewer must approve. A sales team's agent drafts messages a customer answers through the same system. A manager's agent summarizes work that the employee will see and dispute. In each case the agent stands between two people, and its rules govern both. Those rules include when it asks, when it proceeds alone, and which decisions it shows, and the vendor, not either person, sets and changes them. Using the product documentation and release notes of a widely used coding agent, we show how such rules shift between releases. Settings that keep the agent working without stopping for clarifying questions, and questions that time out so the agent proceeds on its own judgement, change where the human decision points fall. We draw three management practices from this. Decide whose rules govern the agent when its output reaches another person. Decide who sees which outputs and decision points. Monitor vendor changes to the agent's behaviour as you would monitor a policy change. A one-page checklist turns each practice into questions a manager can ask before delegating. + +**Keywords:** agentic AI; delegation; governance; third party; vendor change + +## Citation check +| Claim | Source | Depth | +|---|---|---| +| The agent acts on one communicator's behalf; principal–agent analogy | `dxd_foundation/literature/library/hancock2020ai.md` (p. 4, advance access) | full text | +| Agents push humans out of the loop | `dxd_foundation/literature/library/mitchell2026agents.md` | full text (preprint) | +| "nudges Claude to keep working without stopping for clarifying questions" | `phd1750_spaces_in_stream/sec3/claudecode/permission-modes.md:335` | primary document | +| Unanswered questions time out and the agent proceeds | `phd1750_spaces_in_stream/sec3/claudecode/tools-reference.md:136` | primary document | diff --git a/submissions/misqe_delegation_triads/ARGUMENT.md b/submissions/misqe_delegation_triads/ARGUMENT.md new file mode 100644 index 00000000..121eece0 --- /dev/null +++ b/submissions/misqe_delegation_triads/ARGUMENT.md @@ -0,0 +1,14 @@ +# Argument: delegation that creates a third party + +[DRAFTED by Claude, 26 Oct 2026.] + +**P1.** Delegating work to an AI agent is usually framed as a principal–agent relation: the delegator directs, the agent executes, and oversight closes the gap. In AI-mediated communication, the agent acts on behalf of one communicator. +- *Sources:* `dxd_foundation/literature/library/hancock2020ai.md` (autonomy dimension "similar to the principal–agent relationship", p. 4, advance-access pagination); `dxd_foundation/literature/library/mitchell2026agents.md` (approval fatigue; humans pushed out of the loop). + +**P2.** When the agent's output reaches a second person (a reviewer, a customer or a colleague), and that person also acts on or through the same agent, the agent stops being only the delegator's agent. It becomes a third party whose rules both people face, and the vendor sets and changes those rules. Agent products already change how often they ask, when they proceed alone, and which decisions they surface. +- *Sources (primary product documentation, quoted with line numbers in `../phd1750_spaces_in_stream/sec3/claudecode/`):* + - `permission-modes.md:335`: auto mode "nudges Claude to keep working without stopping for clarifying questions" + - `tools-reference.md:136`: an unanswered question times out and Claude "proceeds on its own judgment" + - `CHANGELOG.md:8561`: "Claude will now ask you questions more often in plan mode" + +**C.** Managers delegating to AI agents should treat delegation as creating a triad when outputs cross to another person. They need three practices: decide whose rules govern the agent, decide who sees which outputs and decision points, and detect when the vendor changes the agent's behaviour. They also need a practitioner checklist for each. diff --git a/submissions/misqe_delegation_triads/FLAGS.md b/submissions/misqe_delegation_triads/FLAGS.md new file mode 100644 index 00000000..22c352f1 --- /dev/null +++ b/submissions/misqe_delegation_triads/FLAGS.md @@ -0,0 +1,11 @@ +# Flags: MISQE + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **MISQE wants evidence.** Research Articles must rest on primary research, and Insights may rest on secondary research. Product documentation is secondary evidence, so the Insight route (1,800–3,500 words) is the realistic one. For a Research Article the author would need interviews with managers who delegate to agents. +2. **Conflict of interest.** The evidence is documentation for Claude Code, and Claude drafted this analysis. `../phd1750_spaces_in_stream/paper/OPEN_ISSUES.md` (G15) already flags this. Disclose it, and have the author read every documentation claim against the files. +3. **No claims about motive or retention.** The documentation supports what changed, not why (OPEN_ISSUES G9). +4. **The documentation changes over time.** Quote with version numbers and dates, and re-check the line numbers against the current files before submission. A send-now key added in v2.1.275 makes some earlier "the user must wait" wording outdated (`RESEARCH_NOTES_2026-10-22.md:94`). +5. **The vignettes** (sales agent, manager's summary agent) are illustrative. Mark them as hypothetical in the paper, or replace them with documented cases. +6. **Overlap with `../jitcar_post_ai_enterprise/`:** there the AI is the employer's evaluator. Here it is the delegator's agent that becomes a third party. Keep "organizational capability" out of this paper. +7. **"Substituting GenAI for agentic AI would not change the substance" is grounds for rejection.** The argument must rest on agent autonomy: proceeding alone, timing out and asking less. diff --git a/submissions/misqe_delegation_triads/OUTLINE.md b/submissions/misqe_delegation_triads/OUTLINE.md new file mode 100644 index 00000000..415a9dce --- /dev/null +++ b/submissions/misqe_delegation_triads/OUTLINE.md @@ -0,0 +1,13 @@ +# Outline: MISQE Insight (1,800–3,500 words), or a Research Article (≤7,500 words) if the author collects primary evidence + +[DRAFTED by Claude, 26 Oct 2026. The Insight route is planned at 3,400 words.] + +| § | Section | Words | Mechanism | Sources | +|---|---|---|---|---| +| 1 | The assistant picture and where it breaks | 500 | Three vignettes where output crosses to another person. | `hancock2020ai` | +| 2 | The agent as a third party | 700 | Whose rules, set by whom, and changed when. | `mitchell2026agents`, `starkvandenbroeck2024` | +| 3 | Evidence: how one agent's decision points moved | 900 | Release notes and documentation, line-cited. Shows what changed, with no claims about motive. | `phd1750_spaces_in_stream/sec3/claudecode/` (permission-modes, tools-reference, CHANGELOG, agent-view) | +| 4 | Three management practices | 800 | Govern, show, monitor. | — | +| 5 | Checklist | 300 | One page of questions. | — | +| 6 | Limits | 200 | One product; documentation, not observed use. | — | +| | **Total** | **3,400** | | | diff --git a/submissions/misqe_delegation_triads/README.md b/submissions/misqe_delegation_triads/README.md new file mode 100644 index 00000000..069e5131 --- /dev/null +++ b/submissions/misqe_delegation_triads/README.md @@ -0,0 +1,9 @@ +# *MISQE*: When Your AI Agent Is Also Your Counterpart's + +[DRAFTED by Claude, 26 Oct 2026, from the call page.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **The full paper is due 1 Mar 2027.** The Insight format is proposed (FLAGS 1). Nothing has been submitted. + +**Angle:** delegation to an AI agent creates a triad when outputs cross to another person. It needs three management practices. + +**Relation to other arms:** uses the primary-document corpus in [`../phd1750_spaces_in_stream/`](../phd1750_spaces_in_stream/) (Claude Code documentation). It is distinct from [`../jitcar_post_ai_enterprise/`](../jitcar_post_ai_enterprise/). diff --git a/submissions/misqe_delegation_triads/cfp/README.md b/submissions/misqe_delegation_triads/cfp/README.md new file mode 100644 index 00000000..69f203c1 --- /dev/null +++ b/submissions/misqe_delegation_triads/cfp/README.md @@ -0,0 +1,25 @@ +# CFP brief: *MIS Quarterly Executive*, "Designing, Managing, and Scaling the Delegation of Work to AI Agents" + +[DRAFTED by Claude, 26 Oct 2026. A working digest, not a substitute for the editors' text.] + +**Read on 26 Oct 2026:** https://aisel.aisnet.org/misqe/contact.html; https://aisel.aisnet.org/misqe/policies.html. + +**Editors:** Nick van der Meulen (MIT), Noel Carroll (University of Galway), Hind Benbya (Western Sydney University), Hope Koch (Baylor University). + +## Dates (verbatim) +- "Special Issue full paper submission deadline: March 1, 2027" +- "First editorial review sent to authors: May 1, 2027"; resubmission "July 1, 2027"; final accepted papers "October 1, 2027"; publication "December 2027". +- The ICIS 2026 workshop stage (abstracts due 1 Sep 2026) has passed. "Workshop participation is not a prerequisite for submission to the Special Issue". + +## Format (verbatim, MISQE policies) +- "The manuscript does not exceed 7,500 words (for Research Articles) / ranges between 1,800 - 3,500 words (for MISQE Insight)." +- Research Articles are based on primary research. Insight articles may rest on secondary research or interviews. +- "Submissions should provide clear, accessible insights for executive and practitioner audiences." +- "All papers must be original and not have been previously published or under consideration for publication in any other journal." +- Mark the submission for the special issue on "Agentic AI". + +## Scope lines (verbatim) +- "Overseeing Agents in Action: how organizations design and operate guardrails, monitoring, audit, compliance, trust, and intervention mechanisms when agents act with limited supervision". +- "Reorganizing Work and Leadership: how structures, roles, routines, and management practices change when agents handle execution or coordination". +- "We discourage purely technical papers without a managerial dimension, as well as studies in which substituting "GenAI", "machine learning", or "automation" for "agentic AI" would not change the paper's substance." +- The editors want "evidence-based perspectives grounded in empirical research and industry practice". diff --git a/submissions/orgsci_agency_triad/ABSTRACT.md b/submissions/orgsci_agency_triad/ABSTRACT.md new file mode 100644 index 00000000..cbc92a9f --- /dev/null +++ b/submissions/orgsci_agency_triad/ABSTRACT.md @@ -0,0 +1,18 @@ +# Abstract: *Organization Science*, "Agency in the Age of AI" + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Due 1 Feb 2027.] + +**Title:** Competence and Standing: The Agency of the Co-opted in AI-Mediated Organizing + +## Abstract +When an AI system evaluates two people and binds both to its verdict, who acts? Platforms and AI-mediated workplaces answer by co-optation. A booking platform ranks a hotel on guests' reviews, a marketplace demotes a seller on buyers' ratings, and an AI tool can draft the appraisal a manager gives an employee. In each case the operator benefits from the parties' agency without treating them as its agents, and accountability is distributed, deflected and denied. Recent theory treats AI not as an agent but as a capability of a system of relations among people and algorithms. We ask what agency remains to the people inside such a system. We argue that the co-opted person's agency has two components that vary independently. Competence is the practical ability to interpret a verdict, to address the other party through the system, and to notice when the system's rule changes. Standing is a recognized right to contest a verdict before someone obliged to answer. The two combine into four positions. With both, agency is effective. With competence alone, the person becomes a moral crumple zone, absorbing blame for verdicts they can read but cannot contest. With standing alone, rights go unused. With neither, the person is simply managed. We develop propositions about how organizations distribute competence and standing among the parties they evaluate, why operators prefer to supply competence (training) over standing (contest), and how asymmetric distributions between the two parties shape whom the system blames. The account relocates agency from the person or the machine to the configuration of competence and standing in the triad. + +**Keywords:** agency; co-optation; algorithmic management; accountability; standing + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Benefits from agency without making them agents; distributed, deflected, denied | `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) | full text | +| AI as a capability of a system of relations, not an agent | `lima_pdw/literature/cards/stelmaszak2026.md` (pp. 336, 341) | full text | +| Moral crumple zone | `dxd_foundation/literature/library/elish2019moral.md` (p. 40) | full text | +| Hotel rankings; seller scores; GenAI appraisal (conjecture) | `lima_pdw/literature/cards/orlikowskiscott2023.md`; `lima_pdw/literature/cards/curchod2020.md`; `lima_pdw/literature/cards/hindsvonkrogh2024.md` | full text | diff --git a/submissions/orgsci_agency_triad/ARGUMENT.md b/submissions/orgsci_agency_triad/ARGUMENT.md new file mode 100644 index 00000000..8f9c63eb --- /dev/null +++ b/submissions/orgsci_agency_triad/ARGUMENT.md @@ -0,0 +1,15 @@ +# Argument: competence and standing in the triad + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** On platforms and in AI-mediated work, the operator acts through the agency of the parties it evaluates. It benefits from their agency without making them its agents, and accountability becomes distributed, deflected and denied. AI is better understood as a capability of a system of relations than as an agent. +- *Sources:* `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14); `lima_pdw/literature/cards/stelmaszak2026.md` (pp. 336, 341); `coordinative_sovereignty/literature/library/cards/lazar2025governing.md` (power between mediatees; section I). + +**P2.** A co-opted person's agency then has two parts that vary independently. The first is competence: the practical ability to interpret the system's verdicts, address the other party through it, and notice when it changes. The second is standing: a recognised right to contest a verdict before someone obliged to answer. +- *Sources:* + - `lima_pdw/literature/cards/bovens2010.md` (virtue, a property of one party, versus mechanism, a relation between parties) + - `lima_pdw/literature/cards/kellogg2020.md` (p. 387: often "no procedure in place" to contest) + - `lima_pdw/literature/cards/cameron2024.md` (choice-based consent) + +**C.** Competence without standing turns the competent person into the moral crumple zone: they absorb blame for verdicts they could read but not contest. Standing without competence goes unused. Organization theory should therefore locate agency in the triad as the joint product of the two, and ask how organizations distribute each to the parties they evaluate. +- *Source:* `dxd_foundation/literature/library/elish2019moral.md` (pp. 40, 49). diff --git a/submissions/orgsci_agency_triad/FLAGS.md b/submissions/orgsci_agency_triad/FLAGS.md new file mode 100644 index 00000000..8fa532b1 --- /dev/null +++ b/submissions/orgsci_agency_triad/FLAGS.md @@ -0,0 +1,13 @@ +# Flags: *Organization Science*, Agency in the Age of AI + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The INFORMS PDF is unread.** Length, format and prior-publication rules are unknown. A newer PDF version exists (`...post-1789996319620.pdf`). +2. **Overlap with the OT paper is medium to high.** "Co-optation is binding without consent" is the OT paper's definition. This paper's contribution is **standing**, and the competence × standing configuration. The OT paper uses "standing" only to mean rank. The three practices appear here only as a summary, cited as under review (anonymized). +3. **Overlap with `../jms_organizing_truth_arbiter/`:** this paper owns consent, agency and standing. JMS owns facts, records and truth. Do not use "truth" or "record" here. +4. **Overlap with `../exetasis_sovereignty/`:** this paper stays at the organizational level (standing within the firm or platform). Exetasis owns democratic sovereignty. +5. **Pettit (non-domination) has no read card.** The coordinative_sovereignty cards are search-level or abstract-level. Do not cite Pettit until read. +6. **Emirbayer & Mische, Murray et al. (2021, conjoined agency) and Faraj have no cards.** Murray is a guest editor. Card *Conjoined agency* before writing §2. +7. **Hinds & von Krogh** is conjecture, so write "may". +8. **Stelmaszak et al.** reject "counterpart" as entity-view vocabulary (p. 351). Engage this, because the triad names a counterpart. +9. **Accelerated review, many desk rejections:** the paper must be complete and sharp at submission. diff --git a/submissions/orgsci_agency_triad/OUTLINE.md b/submissions/orgsci_agency_triad/OUTLINE.md new file mode 100644 index 00000000..e8d34d56 --- /dev/null +++ b/submissions/orgsci_agency_triad/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Competence and Standing (*Organization Science*; planned at about 12,000 words; confirm the limit) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 1,200 | Asks who acts when a system judges two people. | `starkvandenbroeck2024`, `stelmaszak2026` | +| 2 | Agency under co-optation | 2,000 | Leveraging without delegating, AI as a relational capability, and power between mediatees. | `starkvandenbroeck2024`, `stelmaszak2026`, `lazar2025governing`, `anthony2023` (abstract) | +| 3 | Two components | 2,200 | Competence (the three practices, summarised and cited as under review) and standing (a recognised right to contest). Why the two vary independently. | `bovens2010`, `kellogg2020` (p. 387), `hirschman1970` | +| 4 | Four positions | 2,000 | A 2 × 2 table: effective agency, crumple zone, unused rights, managed subject. A published case for each. | `elish2019moral`, `cameron2024` (theory only), `curchod2020`, `oborn2025bypassing` | +| 5 | How organizations distribute competence and standing | 2,200 | Training versus contest. Asymmetric distributions between the two parties. Propositions P1–P5. | `rahman2024taming` (abstract), `waardenburg2022` (abstract) | +| 6 | Discussion | 1,800 | Contributions to agency theory and algorithmic management. Boundary conditions. | — | +| 7 | Conclusion | 600 | — | — | +| | **Total** | **12,000** | | | diff --git a/submissions/orgsci_agency_triad/README.md b/submissions/orgsci_agency_triad/README.md new file mode 100644 index 00000000..d168d178 --- /dev/null +++ b/submissions/orgsci_agency_triad/README.md @@ -0,0 +1,12 @@ +# *Organization Science*, "Agency in the Age of AI": Competence and Standing + +[DRAFTED by Claude, 26 Oct 2026, from the guest editors' posts.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **Due 1 Feb 2027; the portal opens 2 Nov 2026.** Nothing has been submitted. + +**Angle:** the co-opted person's agency is the joint product of competence and standing, and organizations distribute both. + +**Relation to other arms:** +- Spun off from [`../algorithmacy_ot/`](../algorithmacy_ot/). The new contribution is standing. +- Distinct from [`../orgsci_ai_frontiers_methods/`](../orgsci_ai_frontiers_methods/), which is a methods special issue in the same journal. +- Distinct from [`../jms_organizing_truth_arbiter/`](../jms_organizing_truth_arbiter/). diff --git a/submissions/orgsci_agency_triad/cfp/README.md b/submissions/orgsci_agency_triad/cfp/README.md new file mode 100644 index 00000000..6dd6bcf0 --- /dev/null +++ b/submissions/orgsci_agency_triad/cfp/README.md @@ -0,0 +1,27 @@ +# CFP brief: *Organization Science*, "Agency in the Age of AI" + +[DRAFTED by Claude, 26 Oct 2026. The official INFORMS PDF returned HTTP 403. Facts come from the guest editors' AOM community posts and the editor-in-chief's Substack. Read the PDF in a browser before writing.] + +**Read on 26 Oct 2026:** +- https://str.aom.org/discussion/organization-science-special-issue-agency-in-the-age-of-ai (posted by Nelson Phillips, 13 Aug 2026) +- https://omt.aom.org/discussion/organization-science-special-issue (10 Aug 2026) +- https://orgsci.substack.com/p/agency-in-the-age-of-ai (Lamar Pierce, 24 Jul 2026) +- https://orgsci.substack.com/p/the-2027-winter-conference-special (18 Oct 2026) + +**Guest editors:** Paul Leonardi, Alex Murray, Frank Nagle, Nelson Phillips, Juliana Schroeder, Paula Ungureanu, Elisa Villani. + +## Dates (verbatim) +- "Submissions Due: February 1, 2027" +- The portal opens 2 Nov 2026 (the 18 Oct Substack post). +- "we will follow an accelerated review process" and "we will not grant extensions to resubmission deadlines." +- "we anticipate a higher than usual number of desk rejections at the initial screening". + +## Scope (verbatim) +- "This special issue seeks theoretically ambitious work, whether conceptual or empirically grounded, that takes the changing production of action as its starting point." +- "We are not looking for general commentary on digital transformation or broad appeals for more AI research." +- "Accountability. When an automated decision causes harm ... it becomes hard to say who acted." +- "Organization theory and institutions. How are accountability, liability, and competence reassigned when action is distributed and outcomes are contested…" +- "Management and control. How do managers coordinate, supervise, and hold to account ensembles of people and autonomous systems…" + +## Unknown +Length, format, timezone and prior-publication rules (all in the blocked PDF). A newer PDF version is linked from the 18 Oct post. diff --git a/submissions/orgsci_ai_frontiers_methods/ABSTRACT.md b/submissions/orgsci_ai_frontiers_methods/ABSTRACT.md new file mode 100644 index 00000000..778583c8 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/ABSTRACT.md @@ -0,0 +1,10 @@ +# ScholarOne fields: *Organization Science*, AI-Enabled Frontiers + +[DRAFTED by Claude, 26 Oct 2026, v3. The abstract field allows 250 words.] + +**Title:** Computing the Triad: An AI-and-PyPhi Pipeline for Testing Whether Coordinative Infrastructure Is Irreducible + +**Running head:** Computing the Triad + +## Abstract +Organizations increasingly coordinate through infrastructure that sits between parties: dispatch systems, claims engines, review gates, AI agents. Platform research calls these arrangements triadic but settles the claim by description. We present a pipeline that computes it. A language model reads case evidence and drafts rival Boolean encodings of who reads whom and what each party commits. Every rule must quote its evidence verbatim, and every encoding carries a prediction fixed in advance. A second, adversarial prompt searches for the most different readings the evidence permits. PyPhi's implementation of integrated information theory 4.0 then computes each encoding exactly. It returns whether any cut along party lines leaves the structure intact, which parties form the irreducible core, the mediator's position (bottleneck, enricher, capture or bypassed), and the single rule on which rival readings diverge. The surface does not decide the verdict: of 256 mediated three-party structures, 24 are irreducible, and time-series proxies separate the two classes barely above chance. In a worked gig-dispatch case, the AI drafter reproduced a human coder's readings, the adversary found two more, and PyPhi refuted both of its predictions. An open repository reproduces every result with one command. In the development stage we will build an encoding module for transcripts and documents, validated against human coders. We will also build a strategic model mapping structural positions and design levers to consequences for platforms, workers and regulators, and validate it on published cases. diff --git a/submissions/orgsci_ai_frontiers_methods/ARGUMENT.md b/submissions/orgsci_ai_frontiers_methods/ARGUMENT.md new file mode 100644 index 00000000..eb7b0da0 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/ARGUMENT.md @@ -0,0 +1,16 @@ +# Argument: an exact test of triadic coordination as shared research infrastructure + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Organization science has no shared, exact test for whether a coordination form is irreducibly three-party or factors into two-party relations. Studies of platforms and algorithmic management decide this by description. +- *Warrant:* the triad is asserted in that literature, not computed. Curchod et al. name "triadic relationships", and Cameron draws an "Algorithmic Labor Triangle", but neither offers a test. +- *Sources:* `lima_pdw/literature/cards/curchod2020.md`; `lima_pdw/literature/models/cameron2024.md` (the triangle and Table 1); the OT claim-check addendum of 11 Oct 2026 (untracked, outside this repo: the author's `wt-algorithmacy-ot` worktree, `submissions/algorithmacy_ot/manuscript/gdoc_2026-10-10/CLAIM_CHECK_ADDENDUM_2026-10-11.md`) places the "Algorithmic Labor Triangle" at p. 466 and finds "irreducible", "triad" and "dyad" absent from the article. + +**P2.** Integrated information computed over the minimum-information partition gives such a test for a formal model of a coordination form. Φ_MIP = 0 means some party-respecting cut factors the form; Φ_MIP > 0 means none does. The lab has built this as a classifier and run it under an AI-assisted pipeline that fixes hypotheses before computing. +- *Sources (repo files):* + - `org_frontier/classifier/README.md` + - `org_frontier/STRUCTURAL_FINDINGS.md`: in the complete 256-form strict-mediation family at n = 3, 24 forms (9.4%) are triadic (`ci/reproduce.json` checks 303 and 503); P(triadic | mediator reads both) = 15.0%; time-series proxies separate dyadic from triadic "only near chance (rank-AUC ≤ 0.63)" (`org_frontier/proxy_bridge/FINDINGS.md`: 0.563 and 0.629; no CI check yet). + - `org_frontier/protocol/RESEARCH_PROTOCOL.md` + +**C.** Released as open infrastructure, the classifier and pipeline let organization scientists turn a described arrangement into a formal model and get an exact, reproducible verdict on its structure. It also shows that the intuitive rule (the mediator reads both parties) over-calls triads: most forms that meet it still factor. That makes it a principled instrument for exploring which arrangements demand triadic competence. +- *Boundary:* the verdicts are about Boolean models. A validation gap separates them from claims about real organizations (`org_frontier/README.md`, line 19). diff --git a/submissions/orgsci_ai_frontiers_methods/FLAGS.md b/submissions/orgsci_ai_frontiers_methods/FLAGS.md new file mode 100644 index 00000000..a2d14fe4 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/FLAGS.md @@ -0,0 +1,22 @@ +# Flags: *Organization Science*, AI-Enabled Frontiers + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **Due 1 Nov 2026,** with three parts: a proposal, a technical appendix and an anonymized repository. The editors' FAQ says "just send us an email" for slight lateness. Do not rely on that. +2. **Anonymity versus the public repo.** algorithmacy-lab is public under the author's GitHub name. The submission needs an anonymized copy (for example an anonymous mirror) with no author names and no commit history. This is the largest practical obstacle. +3. **Fit is low to moderate.** The call is about AI as a research method, not AI as a topic. The proposal only fits if the AI-assisted pipeline and the case-to-model module are the contribution. A paper about algorithmacy does not fit. +4. **Do not claim live LLM-agent probes.** The lab's AI and multi-agent arc uses "in-silico Boolean / tabular / small neural proxies (not live LLM calls...)" (`org_frontier/AI_MULTIAGENT_ARC.md`). The AI in this proposal is the research pipeline (`question_pipeline.js` uses `agent()` calls) and the proposed case-to-model module. +5. **Numbers:** 24 of 256 triadic is in `ci/reproduce.json` (checks 303 and 503). The proxy rank-AUCs (0.563 and 0.629) are in `org_frontier/proxy_bridge/FINDINGS.md` and have **no CI check**. Add one before the appendix cites them. The best-threshold accuracy for the whole-minus-sum proxy is 0.778, so "near chance" refers to rank-AUC only. +6. **Dissertation spine:** Φ is presented as a principled instrument. Nothing here calls it decorative or unnecessary. The limits stated are encoding sensitivity and the validation gap. +7. **The INFORMS PDF was not read.** Format details come from the editors' Substack posts. + +8. **Re-run on 26 Oct 2026** [DRAFTED by Claude]. Census: 24/256; 24/160 = 15.0%; 0/96. Proxy rank-AUCs: 0.563 and 0.629. Controls: pass. The Q43 H1 and H3 checks pass. The anonymized package reproduces all of these standalone. The R1 conditional rates and R2 proxy statistics still have no dedicated `ci/reproduce.json` check. +9. **Correction** [DRAFTED by Claude]. `question_pipeline.js` has no human-approval gate between stages. The researcher chooses the question and accepts or rejects the output. The summary that said "with a human approving each stage" was corrected, and the proposal states human approval only for the *proposed* case-to-model module. +10. **The anonymized repository** is a scrubbed copy: no names, no lab or institution, no git history, and local paths, PR numbers and "dissertation" removed. The call asks for a Git repository, so it must be hosted somewhere anonymous before submission (see the README). +11. **v2, 26 Oct 2026** [DRAFTED by Claude]. The author found v1 did not address the CFP. v2 makes the AI loop the contribution and Φ the oracle. New numbers: 512 hypotheses across 163 parsed questions (77.9% confirmed, 20.5% refuted, 1.6% partial), from an automated parse (`tools/tally_verdicts.py` in the anonymized repo). The refuted share is 17.1% for Q1–Q117 and 24.5% for Q118 onward, which is descriptive only. The audit facts come from `org_frontier/CRITICAL_REVIEW_Q111_Q117.md`. Hand-check a sample of the parse before the workshop. +12. **Anonymity** [DRAFTED by Claude]. In the anonymized repo the coined construct name is replaced by "triadic competence", because searching for it would lead to the public lab. A contributor's workflow log (Q219) is removed. +13. **Proprietary model** [DRAFTED by Claude]. The loop runs on a commercial model today. The FAQ asks for infrastructure others can access, and the proposal commits to an open-weight configuration. +14. **v3, 26 Oct 2026** [DRAFTED by Claude]. Rewritten after the author ruled that v2 missed the CFP. Thesis wording for the author's ruling: the author said organizations will "rely exclusively on irreducible triadic infrastructure"; the proposal says "increasingly coordinate through such infrastructure", because the repo's census finds most mediated structures factor (232/256). Restore "exclusively" if wanted. +15. **Anonymity dropped** [DRAFTED by Claude]. The CFP PDF does not require it. The repo is public under the author's name (MIT; CITATION.cff with ORCID). +16. **Example predictions** [DRAFTED by Claude]. The examples' predictions were written knowing the lab results, and the repo README says so. Only the logged AI runs are blind. +17. **The [REPOSITORY URL] placeholder** appears in PROPOSAL.md and TECHNICAL_APPENDIX.md and must be replaced before the PDFs are rebuilt. diff --git a/submissions/orgsci_ai_frontiers_methods/OUTLINE.md b/submissions/orgsci_ai_frontiers_methods/OUTLINE.md new file mode 100644 index 00000000..c310bfe9 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: proposal package (*Organization Science*, AI-Enabled Frontiers) + +[DRAFTED by Claude, 26 Oct 2026. Proposal of 1,500–2,000 words plus references; technical appendix; repository.] + +| Part | Words | Content | Sources | +|---|---|---|---| +| 1. The problem | 300 | "Triadic" is asserted in the literature and never tested. Two published cases are described as three-party. | `curchod2020`, `lima_pdw/literature/models/cameron2024.md`, `stark2020algorithmic` (p. 53) | +| 2. What we have done | 450 | The classifier, the 256-form census, the proxy failure, and the six-stage AI-assisted pipeline with human gates. | `org_frontier/classifier/README.md`, `org_frontier/STRUCTURAL_FINDINGS.md` (lines 18–20, 46–48), `org_frontier/protocol/RESEARCH_PROTOCOL.md` | +| 3. What we will do | 450 | Release the classifier with a model-writing guide. Build a case-to-model module in which AI drafts candidate models and humans approve them. Run it on 10–20 published platform cases. Report verdicts with sensitivity to encoding. | `org_frontier/field/README.md` (version-0 protocol, ten mocks) | +| 4. Key risks | 300 | (a) Encoding sensitivity: Φ magnitude is not a scale, so only the binary verdict is read. (b) The model-to-organization validation gap. (c) Computational cost beyond five nodes. | `org_frontier/classifier/README.md`; `org_frontier/README.md` line 19 | +| 5. If successful | 300 | Structural claims become exact and comparable. Theory can ask which arrangements demand a triadic competence. | — | +| Technical appendix | — | Definitions (Φ over the MIP versus the complete cut), validation on the lab's own controls, and runtime. | `ci/reproduce.json` checks | +| Repository | — | An anonymized fork of `org_frontier/classifier/` and `protocol/`, with no author names and no lab history. | — | +| **Total** | **1,800** | | | diff --git a/submissions/orgsci_ai_frontiers_methods/PROPOSAL.md b/submissions/orgsci_ai_frontiers_methods/PROPOSAL.md new file mode 100644 index 00000000..81da412e --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/PROPOSAL.md @@ -0,0 +1,85 @@ +# Computing the Triad: An AI-and-PyPhi Pipeline for Testing Whether Coordinative Infrastructure Is Irreducible + +[DRAFTED by Claude, 26 Oct 2026, v3, planned against the CFP PDF only. Stage 1 proposal for *Organization Science*, "AI-Enabled Frontiers in Organizational Science" (due 1 Nov 2026). Every number reproduces from the submission repository with `python reproduce.py`; the repository URL replaces [REPOSITORY URL] once it is created. v1 and v2 are retired; see git history.] + +[Current version: https://docs.google.com/document/d/1fhM_C7pz6e3sR7aXY3_89oys06B-UyK9Vyf93bt1zFA/edit — the author's Google Doc supersedes this draft.] + +## The structural fact strategy needs and cannot see + +Rahman's (2021) freelancers worked under a score that moved without notice, and among the things they could not decipher was who had moved it, the client or the platform. Curchod, Patriotta, Cohen and Neysen (2020) describe eBay's customer reviews as enacting triadic relationships among the platform operator, buyers and sellers, where a dyadic exchange had stood. Stark and Pais (2020) give the platform a triangular geometry in which the owner co-opts providers and users alike, and Cameron (2024) draws an algorithmic labor triangle. Organizations increasingly coordinate through such infrastructure: a dispatch system between driver and rider, an engine between claimant and adjuster, an AI agent between two firms. The literature calls these arrangements triadic by description. Nobody computes it. + +The surface does not settle it. In the complete family of 256 three-party structures in which the outer parties have no direct link and must coordinate through a mediator, 24 are irreducibly triadic. The intuitive test, that the mediator reads both parties, is necessary but nowhere near sufficient: it holds for 160 structures, of which 24 are triadic, and fails for 96, of which none is. Most arrangements with a platform in the middle still factor into pairs. + +Irreducibility is a strategic fact. It says who is bound to whom, who can leave, and what a redesign would change. We have built a method that computes it exactly for any small coordinative structure written as a model of who reads whom and what each party commits. We have not yet built the model of what to do about it. This proposal describes the method, shows it working, and lays out how we will build the strategic model on top of it. + +## The pipeline: what AI does, what PyPhi computes, what the researcher decides + +1. **Bound and name.** The researcher fixes one recurring act (one match, one claim, one merge), names the parties as nodes, and states what each party's active bit means and what that binary collapses. *Built, as a written field protocol with a coding scheme, interview guide and ethics note.* +2. **Encode.** A language model drafts rival Boolean encodings of the arrangement, one per defensible reading. Each party's next state is an expression over the parties it reads, such as `P' = D and R`, and every rule must quote the passage it rests on. For the mediator, the model must also say whether the evidence shows it *committing* on a joint condition, *conveying* one party's signal, or *storing* inputs for someone else. A second, adversarial prompt then searches for the most different readings the evidence still permits: commit versus convey, interchangeable parties, and outside options. Every encoding carries a predicted verdict fixed before anything is computed. A checker rejects any rule whose quotation does not appear verbatim in the source. *Built as prompts, a schema and the checker, and run on two cases; ingestion of interview transcripts at scale is planned.* +3. **Compute.** For each encoding the rules become a state-by-node transition matrix and a connectivity matrix. PyPhi's implementation of integrated information theory 4.0 (Albantakis et al., 2023; Mayner et al., 2018) evaluates every reachable state and returns system integrated information Φ over the minimum-information partition (the cut that damages the cause–effect structure least), together with the major complex, the subset of parties that forms the irreducible core. If Φ is zero in every state, some cut along party lines leaves the structure intact and the arrangement is *dyadic*. If Φ is positive in any state, no cut does, and it is *triadic*. *Built.* +4. **Validate first.** Before reading any verdict, the instrument must reproduce its controls: a structure with a decoupled party reads dyadic, and a fully coupled one reads triadic. If either fails, every script refuses to run. *Built.* +5. **Report the spread.** The pipeline computes every approved encoding, reports the Φ gap and core overlap between rival readings, and names the single rule whose change flips the verdict. A verdict that holds across every defensible encoding is robust. One that does not is contested, and the flipping rule is the fact fieldwork must settle. *Built.* +6. **Read the structure.** Three readings follow from the verdict. *Who is bound* is the major complex. *The mediator's position* comes from a lesion test: a bottleneck carries the coordination; an enricher adds to a coordination that survives its removal; capture binds the mediator to some parties and leaves others out; a bypassed mediator sits outside a core the parties form themselves. *Intrinsic or contingent* comes from restoring a forbidden direct tie: an intermediary that does integrative work stays in the core, while one held there only by a constraint on bypassing it drops out. *Built.* + +The division of labour is the design. The model proposes readings at a breadth no single coder sustains; PyPhi decides each exactly; the researcher bounds the act, approves encodings, fixes predictions and interprets the result for strategy. + +## A walk-through: gig dispatch + +The repository's first worked example applies the pipeline to a stipulated description of one dispatch event. The platform's manual says the system "assigns a trip only when an eligible driver is available and a rider has confirmed the request." A driver says "It's a suggestion, it follows what I do." The manual adds that "any eligible driver in range may be assigned," and some riders text their regular driver when the app is slow. + +The platform's account encodes as `D' = P, P' = D and R, R' = P`. PyPhi returns Φ = 2.0 at the state where all three are active, with all three parties in the core and the platform a bottleneck. Restoring a direct driver–rider tie leaves the platform in the core, so its place is intrinsic. The driver's account changes one rule, `P' = D`, and the structure factors (Φ = 0). The pipeline reports that this single rule carries the whole disagreement, so the question a study must answer is whether dispatch commits or relays. Pooling drivers, `P' = (D1 or D2) and R`, also factors the structure and removes the individual driver from the core, leaving platform and rider bound. + +Holding the platform's rule fixed and varying outside options gives the strategic reading. With both sides captive, the platform is a bottleneck. With a complementary direct tie, it is an enricher. With a captive driver and a mobile rider, it captures the driver; give the driver the option instead and the lock-in moves to the rider. Only symmetric options leave the platform outside the core. For a platform, a drivers' collective or a regulator, these are the structural consequences of each lever. Which lever to pull in which circumstance is the strategic model we will build. + +The AI step ran on this case. Given the description alone, the drafter reproduced the two rival readings a human coder wrote, with every quotation verbatim. The adversary then found the driver pool and the substitute direct tie, and declined a complementary tie that the text does not support. Both of its predictions failed when computed. It expected the pooled structure to stay triadic, but it factors. It expected the direct-tie structure to factor, but it stays irreducible with the platform bypassed. The model found the readings, and PyPhi decided them. On a second case, insurance-claim routing, the drafter declined to write rules for four of five designs because the description did not specify them, and the checker rejected the file. The gate worked as intended. + +## What we have done + +The repository ([REPOSITORY URL]) holds the pipeline as installable code, four worked examples (gig dispatch, insurance-claim routing, a code-review gate, and a dealer contrasted with a clearinghouse), the logged AI runs, the field protocol, 19 tests, and one command that reproduces all 17 reported results in about two minutes on a laptop. + +The reference results show why the verdict must be computed rather than read off a diagram, and why its encoding must be searched rather than chosen. Beyond the 256-structure census: +- Two standard time-series proxies for integration separate dyadic from triadic structures at rank-AUC 0.563 and 0.629, barely above chance. +- In ten invented organizations, the computed verdict matched the reading fixed before computing in nine. Four flipped under a second encoding the same story permits. +- An adversary limited to readings the evidence permits flipped the point verdict on all 101 contested accounts it attacked. + +Encoding sensitivity is the method's central fact, and the pipeline reports it instead of hiding it. The outside-option law behind the walk-through holds on 60 of 60 random configurations. One outside contributor, working on their own machine with an AI assistant, took a new question (who decides an insurance claim?) through the laboratory's protocol to a reproducible result. + +## What we hope to do + +Stage 2 (January–May 2027) would build three things. + +- **The encoding module, before the February workshop.** It will ingest interview transcripts, documents and logs; attach a locator to every rule; run the drafter and the adversary on open-weight as well as commercial models; and put a researcher's approval between drafting and computing. We will measure agreement between the AI drafter and two human coders against the coding scheme's bar (Krippendorff's α of 0.80), so workshop participants can run their own cases. +- **The strategic model (February–April).** An actor-indexed table maps a structural position (verdict, core, mediator regime, intrinsic or contingent) and a design lever (add or remove a binding, break or restore liveness, require all parties or allow substitutes, open or close a bypass, change an outside option) to its consequence: who stays bound, who is freed, whether the mediator survives. Every cell is a computation the pipeline can already run. The model is structural and makes no claims about economic value. +- **Validation on real cases (March–May).** We will encode 10 to 15 published platform and algorithmic-management cases from their own text and report, for each, the verdict under every defensible encoding and the detail that would decide between them. The interview packet for gig dispatch is complete and will be fielded once ethics approval is in hand. + +The Stage 3 article would report the pipeline, the walk-through, the table of published cases, the lever table and the agreement between AI and human coders. If development stops short, a letter reporting encoding sensitivity alone would still serve anyone who computes structure from text. + +## Key risks + +- **Encoding sensitivity.** This is the main risk. A verdict is exact for its encoding, not for the organization. The design response is to search the space of encodings, require human sign-off, and count a case as robust only when every defensible encoding agrees. +- **The validation gap.** A verdict on a model says nothing yet about how workers experience an arrangement; that needs fieldwork. +- **Scale.** Exact Φ is practical to about ten binary parties. Most coordination units have three to five. +- **The model.** The drafter currently runs on a commercial model, and an open-weight configuration is planned. The strategic model does not yet exist. +- **Ethics.** Verdicts concern arrangements, not persons. Field cases proceed only under approval. The same analysis that shows a platform how capture works shows workers and regulators which move dissolves it. + +## How organizational science would change + +Structural claims about triadic infrastructure would become computed, comparable and contestable. A study would publish its encodings, and a reader could rerun the verdict under a rival reading. Typologies built on description could be tested against one standard: Thompson's (1967) ordering of pooled, sequential and reciprocal interdependence collapses at its top step when computed. Strategy would gain a vocabulary of levers with exact structural consequences. Daft and Lewin (1990) asked organization studies to break out of the normal-science straitjacket. Borrowing an exact measure of irreducibility from another field, with AI doing the reading that makes it usable, is one way out, and the repository is where anyone can start. + +## References + +Albantakis, L., Barbosa, L., Findlay, G., Grasso, M., Haun, A. M., Marshall, W., Mayner, W. G. P., Zaeemzadeh, A., Boly, M., Juel, B. E., Sasai, S., Fujii, K., David, I., Hendren, J., Lang, J. P., & Tononi, G. (2023). Integrated information theory (IIT) 4.0: Formulating the properties of phenomenal existence in physical terms. *PLOS Computational Biology*, 19(10), e1011465. https://doi.org/10.1371/journal.pcbi.1011465 + +Cameron, L. D. (2024). The making of the "good bad" job: How algorithmic management manufactures consent through constant and confined choices. *Administrative Science Quarterly*, 69(2), 458–514. + +Curchod, C., Patriotta, G., Cohen, L., & Neysen, N. (2020). Working for an algorithm: Power asymmetries and agency in online work settings. *Administrative Science Quarterly*, 65(3), 644–676. https://doi.org/10.1177/0001839219867024 + +Daft, R. L., & Lewin, A. Y. (1990). Can organization studies begin to break out of the normal science straitjacket? An editorial essay. *Organization Science*, 1(1), 1–9. + +Mayner, W. G. P., Marshall, W., Albantakis, L., Findlay, G., Marchman, R., & Tononi, G. (2018). PyPhi: A toolbox for integrated information theory. *PLOS Computational Biology*, 14(7), e1006343. https://doi.org/10.1371/journal.pcbi.1006343 + +Rahman, H. A. (2021). The invisible cage: Workers' reactivity to opaque algorithmic evaluations. *Administrative Science Quarterly*, 66(4), 945–988. https://doi.org/10.1177/00018392211010118 + +Stark, D., & Pais, I. (2020). Algorithmic management in the platform economy. *Sociologica*, 14(3), 47–72. https://doi.org/10.6092/issn.1971-8853/12221 + +Thompson, J. D. (1967). *Organizations in action: Social science bases of administrative theory*. McGraw-Hill. diff --git a/submissions/orgsci_ai_frontiers_methods/README.md b/submissions/orgsci_ai_frontiers_methods/README.md new file mode 100644 index 00000000..8cc2b377 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/README.md @@ -0,0 +1,27 @@ +# *Organization Science*, AI-Enabled Frontiers: an exact test for triadic coordination + +[DRAFTED by Claude, 26 Oct 2026, from the editors' posts. The angle is a proposal for the author to rule on.] + +**Current version:** the up-to-date proposal is the author's Google Doc: . The Markdown files here are Claude's v3 draft and the review against that doc; where they differ, the Google Doc governs. + +**Status:** v3 drafted 26 Oct 2026 and not yet submitted. It was planned against the CFP PDF only, with Fable planning the writing. The proposal (1,984 words) presents the AI + PyPhi pipeline that tests a coordinative structure for irreducibility, and the strategic model as Stage 2. The submission repository `irreducibility-pipeline` is built and committed locally at `~/iit-playground/irreducibility-pipeline` (tag `v0.1-stage1`; 19 tests and 17 reproduce checks pass in a fresh venv) and is not yet pushed to GitHub. v1 and v2 are retired. **Due 1 Nov 2026.** + +**The call:** a special issue that develops AI-enabled research infrastructure, with a workshop at Duke on 19–20 Feb 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the editors) | +|---|---| +| Proposal, technical appendix and repository | **1 Nov 2026** | +| Development workshop | 19–20 Feb 2027 (date may change) | +| Final papers (2,500–4,500 words) | Summer 2027 | + +**Angle (v3):** exactly how AI and PyPhi test a coordinative structure for irreducibility, and why that matters for strategy as organizations rely on triadic infrastructure. The GitHub repo is the prototype. + +**Relation to other arms:** draws on `org_frontier/` (classifier, protocol, structural findings). It is distinct from [`../orgsci_agency_triad/`](../orgsci_agency_triad/), which is a theory paper for a different *Organization Science* special issue. + +| File | What it is | +|---|---| +| `ABSTRACT.md` | 250-word proposal summary | +| `ARGUMENT.md` | P1, P2, C | +| `OUTLINE.md` | Proposal package plan | +| `FLAGS.md` | Anonymity, fit and number checks | +| `cfp/` | Digest; `source/` is gitignored | diff --git a/submissions/orgsci_ai_frontiers_methods/TECHNICAL_APPENDIX.md b/submissions/orgsci_ai_frontiers_methods/TECHNICAL_APPENDIX.md new file mode 100644 index 00000000..5d2dc7cb --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/TECHNICAL_APPENDIX.md @@ -0,0 +1,121 @@ +# Technical Appendix: Computing the Triad + +[DRAFTED by Claude, 26 Oct 2026, v3. Companion to the proposal. Every result below reproduces from the submission repository ([REPOSITORY URL]) with `python reproduce.py` (17 checks, about two minutes) and `pytest` (19 tests).] + +[Current version: https://docs.google.com/document/d/1fhM_C7pz6e3sR7aXY3_89oys06B-UyK9Vyf93bt1zFA/edit — the author's Google Doc supersedes this draft.] + +## A. Encoding a coordinative structure + +A structure with *n* parties is a deterministic Boolean network on *n* nodes. Node *i* holds party *i*'s state, and a rule sets its next state from the current states of the parties it reads. Rules are written as expressions over party labels using `and`, `or`, `not`, `xor`, `0` and `1` (`irreducibility/rules.py`). The parser rejects anything else, so an encoding cannot execute code. + +An encodings file (`encode/SCHEMA.json`) holds one structure and several encodings. Each encoding gives: +- **role:** `rival` for a competing reading of the same arrangement, `design` for a deliberate change to it; +- **rules:** one per party; +- **evidence:** for every rule, a source, a locator and a verbatim quotation; +- **alternative:** for the mediator's rule, the reading the evidence does not rule out, and why; +- **predicted:** the verdict and core, with a reason, fixed before computing; +- **bypass (optional):** a direct tie for the contingency test. + +Strict mediation, for example: + +| Party | Rule | Reading | +|---|---|---| +| Driver D | D′ = P | acts on the platform's signal | +| Platform P | P′ = D and R | commits only when both parties are in | +| Rider R | R′ = P | responds to what the platform shows | + +## B. The computation (`irreducibility/core.py`) + +1. **Transition matrix.** `tpm_from_rules` builds the 2ⁿ × n state-by-node matrix, little-endian (index 0 = first party), as PyPhi expects. `cm_from_rules` builds the connectivity matrix by a flip test: cm[i, j] = 1 iff flipping node *i* can change node *j*'s next state. +2. **Reachable states.** These are states with at least one predecessor under the dynamics; PyPhi analyzes only these. +3. **Φ over the MIP.** For each reachable state, `pyphi.new_big_phi.sia` returns system integrated information over the minimum-information partition (Albantakis et al., 2023), together with the partition itself. +4. **Verdict.** The structure is **triadic** iff Φ > 10⁻⁹ in at least one reachable state, read at the most-integrated state; otherwise it is **dyadic**. The test uses the MIP, not the complete cut. The complete cut severs any coupled pair and would call every coupled dyad a triad. +5. **Major complex.** `pyphi.new_big_phi.maximal_complex` is maximized over reachable states. It names the parties bound into the irreducible core. +6. **Interpretation.** Only the binary verdict and core membership are interpreted. Φ magnitude depends on the encoding and is at most ordinal. + +**Environment.** PyPhi is pinned to commit `b78d0e342` of its IIT-4.0 line (`requirements.txt`), with Python 3.10 or later; the code is tested on 3.12. + +## C. Controls (`python -m irreducibility.validate`; every script refuses to run if they fail) + +| Control or form | Rules (W, S, C) | Expected | Observed | +|---|---|---|---| +| Factoring control | S, W, C | dyadic | dyadic, Φ = 0.000 | +| Irreducible control | S or C, W and C, W xor S | triadic | triadic, Φ = 0.830 | +| Strict-mediation triad | S, W and C, S | triadic | triadic, Φ = 2.000 | +| Chat dyad; dyadic gig model; feedback that breaks liveness | (see code) | dyadic | dyadic, Φ = 0.000 each | +| Gig "false dyad" | not S, W and C, C and not S | triadic | triadic, Φ = 2.000 | + +## D. Structural readings + +- **Mediator position** (`irreducibility/regime.py`): + - *dyadic*: the structure factors. + - *BYPASSED*: triadic, but the mediator is outside the core. + - *CAPTURE*: the mediator is in the core and some party is not. + - *BOTTLENECK*: all parties are in the core, and freezing the mediator at 0 or 1 leaves the others unable to coordinate. + - *ENRICHER*: all parties are in the core, and the others still coordinate with the mediator frozen. +- **Intrinsic or contingent** (`irreducibility/contingency.py`). Restore the forbidden direct tie (the downstream party reads the upstream party) and recompute the core. + - *Contingent*: the intermediary leaves the core. + - *Intrinsic*: it stays, with no loss of Φ. + - *Partial*: it stays, but Φ falls. + - *Reducible*: it was never in the core. +- **Spread across rival encodings** (`irreducibility/spread.py`, `analyze.py`): verdict agreement, the Φ gap, and the Jaccard overlap of the cores. For every pair of rival encodings with different verdicts, the analysis names each single rule whose swap flips the verdict. + +## E. The AI step + +| Component | File | Function | +|---|---|---| +| Schema | `encode/SCHEMA.json` | the format a drafter must produce | +| Drafter | `encode/PROMPT_drafter.md` | rival encodings with verbatim evidence, an alternative and a fixed prediction; commit, convey or store decided from evidence; pools encoded with `or` | +| Adversary | `encode/PROMPT_adversary.md` | the most different readings the evidence still permits on three forces (commit vs convey, substitutability, spectators and outside options), or a statement of which passage rules each one out | +| Gate | `irreducibility/encode_check.py` | rejects a missing rule, a rule with no evidence, a quotation not found verbatim in the source, a mediator rule with no alternative, or a missing prediction; never computes Φ | + +**Logged runs.** Claude Sonnet 5.5 ran on 26 Oct 2026 with the committed prompts and the case description only, with no tools and no access to the human encodings. + +| Run | Output | Result | +|---|---|---| +| Gig dispatch, drafter | `examples/01_gig_dispatch/ai_draft.json` | admissible; two rival encodings identical to the human encodings A and B; both predictions held; the driver pool was missed | +| Gig dispatch, adversary | `examples/01_gig_dispatch/ai_draft_plus_adversary.json` | added the driver pool (computed dyadic, core {P,R}; predicted triadic, so the prediction **failed**) and the substitute direct tie (computed triadic, core {D,R}, platform bypassed; predicted dyadic, so it **failed**); declined a complementary tie as unsupported | +| Insurance claim, drafter | `examples/02_insurance_claim/ai_draft.json` | left rules unwritten for four of five designs and stated the missing evidence; the gate **rejected** the file (9 problems) | + +## F. Worked examples (`examples/`; stipulated descriptions, not field data) + +The lab's results for these structures were known when the examples were written, so their predictions record the naive reading, not a blind test. + +| Example | Encodings | Results | +|---|---|---| +| 01 Gig dispatch | 3 rival, 5 design | Platform commits: triadic, Φ = 2.0, core {D,P,R}, bottleneck, intrinsic. Driver's relay: dyadic. Pooled drivers: dyadic, core {P,R}. Flip rule: P: `D and R` → `D`. Outside options O1–O5: bottleneck, enricher, capture {D,P}, capture {P,R}, bypassed {D,R}. | +| 02 Insurance claim | 5 design | Pipe: triadic, core {C,E,A}. Threshold auto-approval: core {C,E}, capture. Fraud loop with learning: core {E,A}. Loop without learning: core {C,A}, engine bypassed. Frozen claimant: dyadic. 2 of 5 naive predictions held. | +| 03 Code-review gate | 2 rival | Maintainers wait for CI: triadic, Φ = 1.0, core {A,M}, gate bypassed. Maintainers approve on the diff alone: core {A,G,M}, gate enricher. | +| 04 Dealer vs clearinghouse | 2 design | Dealer under a ban on direct sales: core {U,I,V}, contingent (Φ lost when bypassed: 2.0). Clearinghouse: intrinsic (Φ lost: 0.0). | + +## G. Reference results (`studies/`) + +| Check | Result | +|---|---| +| `census` | 256 strict-mediation structures (4 worker rules × 16 mediator rules × 4 counterpart rules). 24 are triadic; 24/160 where the mediator reads both parties and 0/96 otherwise. | +| `field-mocks` | Ten invented organizations: the verdict matched the reading fixed before computing in 9/10 (support-ticket triage read triadic against a dyadic prediction). 4 flip under a second defensible encoding. | +| `mediator-in-core`, `enricher-regime` | Over 500 random mediated forms: dyadic 44.4%, bypassed 18.8%, capture 28.2%, bottleneck 2.6%, enricher 6.0%. A platform's core is itself plus the parties without an outside option (60/60). | +| `contingent-irreducibility` | The bypass test realizes the intrinsic / contingent / partial / reducible taxonomy. | +| `design-operations` | Three levers (binding, liveness, requirement), each with a build operation and an exact inverse break. | +| `thompson-ordering` | Pooled Φ = 0, sequential Φ = 2.0, reciprocal Φ = 2.0: the top step of Thompson's ordering collapses. | +| `adversarial-coding` | An adversary limited to evidence-permitted readings flips the point verdict on 101/101 contested accounts; a coder-agreement interval brackets the adversarial reading in only 22. | +| `proxies` | ΦID Φ_R and whole-minus-sum Φ from simulated trajectories separate dyadic from triadic at rank-AUC 0.563 and 0.629. | + +## H. Planned specifications (Stage 2) + +- **Encoding module.** + - Input: transcripts, documents and logs, with locators. + - Process: drafter and adversary on an open-weight and a commercial model; a researcher's approval recorded per encoding. + - Agreement: AI drafter against two human coders, Krippendorff's α per rule and per bit, target ≥ 0.80. +- **Strategic model.** + - Rows: structural position (verdict, core, regime, intrinsic or contingent), indexed by actor (platform, worker, counterpart, regulator). + - Columns: design levers (add or remove binding; break or restore liveness; require all parties or allow substitutes; open or close a bypass; change an outside option). + - Cells: the computed consequence (who stays bound, who is freed, whether the mediator survives). + - Scope: structural only, with no claims about economic value. +- **Validation.** 10–15 published platform and algorithmic-management cases encoded from their own text. For each case, the verdict under every defensible encoding and the detail that would decide between them. The gig-dispatch interview packet (`protocol/`) will be fielded under ethics approval. + +## References + +Albantakis, L., et al. (2023). Integrated information theory (IIT) 4.0: Formulating the properties of phenomenal existence in physical terms. *PLOS Computational Biology*, 19(10), e1011465. https://doi.org/10.1371/journal.pcbi.1011465 + +Mayner, W. G. P., Marshall, W., Albantakis, L., Findlay, G., Marchman, R., & Tononi, G. (2018). PyPhi: A toolbox for integrated information theory. *PLOS Computational Biology*, 14(7), e1006343. https://doi.org/10.1371/journal.pcbi.1006343 diff --git a/submissions/orgsci_ai_frontiers_methods/cfp/README.md b/submissions/orgsci_ai_frontiers_methods/cfp/README.md new file mode 100644 index 00000000..80618399 --- /dev/null +++ b/submissions/orgsci_ai_frontiers_methods/cfp/README.md @@ -0,0 +1,28 @@ +# CFP brief: *Organization Science*, "AI-Enabled Frontiers in Organizational Science" + +[DRAFTED by Claude, 26 Oct 2026. The official INFORMS PDF returned HTTP 403. Everything here comes from the editors' own Substack posts. Confirm against the PDF in a browser.] + +**Read on 26 Oct 2026:** +- https://orgsci.substack.com/p/ai-enablement-for-better-research (Lamar Pierce, 6 Jul 2026; reproduces the call) +- https://orgsci.substack.com/p/the-portal-is-open (Hasan & Gartenberg, 9 Oct 2026; FAQ transcript) +- Blocked: https://pubsonline.informs.org/pb-assets/filesorsc/CFP_AI-Enabled%20Frontiers%20in%20Organizational%20Science-1785685920943.pdf + +**Editors:** Claudine Gartenberg, Sharique Hasan, Lamar Pierce, Christopher Bail, Hengchen Dai, Oliver Hauser, Hatim Rahman, Dennis Zhang. + +## Stages (verbatim) +- "Stage 1: Submit research proposal and package (Deadline: November 1, 2026)" +- "Stage 2: Engage in development process (January-May, 2027)". The workshop is at Duke on 19–20 Feb 2027, per the 9 Oct post. The July post said 27–28 Feb. +- "Stage 3: Finalization of research projects (Summer, 2027)" +- Final papers are "Nature-style ... between 2,500 and 4,500 words." + +## Package (verbatim) +- "A short proposal: approximately 3-5 pages / 1,500–2,000 words plus references, which should describe the core intended scientific contribution. This should include a description of what you have done, what you hope to do, the key risks you face in fully executing your vision, and how organizational and social science would change if you are successful." +- "Technical appendix or supplement". +- "Research artifact: prototype and repository. Authors should provide a Git repository or equivalent artifact that includes a prototype or proof of concept for the work." +- Anonymity: "make sure that whatever artifacts you upload to that repo are anonymous"; "Make sure that your proposal does not have author names in it". +- Portal: mc.manuscriptcentral.com/orgsci, "Special Issue: AI-Enabled Frontiers". + +## Scope (verbatim) +- Contribution types include "AI-enabled research loops under human direction", "reusable research infrastructure", "new forms of measurement" and "Synthetic social systems". +- "what we do not want is more obvious: conventional full-length papers with AI-related content or typical application." +- "Can this be theoretical? Yes, we welcome work that uses AI to build theory, provided the process is transparent and the theoretical contribution is clear." diff --git a/submissions/perspsych_measuring_algorithmacy/ABSTRACT.md b/submissions/perspsych_measuring_algorithmacy/ABSTRACT.md new file mode 100644 index 00000000..efbddd6f --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/ABSTRACT.md @@ -0,0 +1,19 @@ +# Abstract: *Personnel Psychology*, "The Generative Workplace" (a planned study; results not yet available) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. **No data exist yet.** Bracketed text marks results that are to come. Window: 1–27 Aug 2027.] + +**Title:** Algorithmacy at Work: Measuring the Competence to Coordinate With Colleagues Through Generative AI + +## Abstract +Generative AI now stands between colleagues. It can draft appraisals, screen submissions and gate work through automated checks, so much of the relational experience of generative AI at work is the experience of reaching another person through a system that evaluates both. Existing measures of competence with algorithms score a person's relation to the system alone. A validated scale of platform workers' algorithmic competency asks about understanding, embracing, leveraging and remediating algorithmic management. None of its items concerns coordinating with the human counterpart. We develop and test the Algorithmacy Competence Scale, a 12-item measure with three facets. The first is counterpart inference, which means reconstructing what a hidden counterpart wants from outcomes. The second is signal compression, which means expressing intent in the few signals the system accepts. The third is rule-change tracking, which means noticing changes the system makes without announcement. In a three-wave panel of a 16-week developer cohort, in which work passes peer review, votes and automated checks, we test the three-factor structure, measurement invariance across waves, growth over the program, and discriminant validity against the algorithmic-competency scale. The hypotheses and decision rules were fixed before data collection. [RESULTS TO COME: factor structure, invariance, growth slope, latent correlation with algorithmic competency.] We discuss what the measure adds to research on the generative workplace and its limits, including the small single-cohort sample. + +**Keywords:** algorithmacy; generative AI; scale development; coordination; longitudinal measurement + +## Citation check +| Claim | Source | Depth | +|---|---|---| +| Four dimensions of algorithmic competency; no counterpart item | `lima_pdw/literature/cards/zhou2025apjhr.md` | full text | +| GenAI in appraisal and screening (conjecture) | `lima_pdw/literature/cards/hindsvonkrogh2024.md` (pp. 4, 6) | full text | +| Scale facets and items | `org_frontier/survey/cohort_algorithmacy/codebook.md` | repo | +| Design, waves and setting | `org_frontier/survey/cohort_algorithmacy/STUDY.md` (lines 160–178) | repo | +| Hypotheses fixed before data | `org_frontier/survey/cohort_algorithmacy/PRE_REGISTRATION.md` | repo (git history only) | diff --git a/submissions/perspsych_measuring_algorithmacy/ARGUMENT.md b/submissions/perspsych_measuring_algorithmacy/ARGUMENT.md new file mode 100644 index 00000000..c6b2e83d --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/ARGUMENT.md @@ -0,0 +1,12 @@ +# Argument: measuring algorithmacy + +[DRAFTED by Claude, 26 Oct 2026.] + +**P1.** Existing measures of competence with algorithms score a person's relation to a system: awareness and knowledge of algorithms, or a platform worker's understanding of, embracing of, leveraging of and remediating of algorithmic management. None of their items refers to coordinating with another person through the system. +- *Sources:* `lima_pdw/literature/cards/zhou2025apjhr.md` (Table 2; "No item measures a capacity to coordinate with the human counterpart", card); `lima_pdw/literature/cards/dogruel2022.md` (abstract depth). + +**P2.** Generative AI increasingly stands between colleagues. It drafts reviews, filters submissions and gates work through automated checks, so the relational experience of GenAI at work includes coordinating with others through a system that evaluates both. In the lab's developer cohort, work passes a peer-review-and-vote gate and automated checks. +- *Sources:* `lima_pdw/literature/cards/hindsvonkrogh2024.md` (pp. 4, 6; conjecture); `org_frontier/survey/cohort_algorithmacy/STUDY.md` (setting). + +**C.** A valid measure of that experience needs items that refer to the counterpart and the system together. The study tests a three-facet scale (counterpart inference, signal compression, rule-change tracking). It tests the facet structure, invariance and growth over a 16-week cohort, and discriminant validity against the APJHR algorithmic-competency scale. +- *Instrument:* `org_frontier/survey/cohort_algorithmacy/codebook.md` (12 items, 3 × 4); hypotheses H1–H4 in `PRE_REGISTRATION.md`, fixed before data collection (no external registry). diff --git a/submissions/perspsych_measuring_algorithmacy/FLAGS.md b/submissions/perspsych_measuring_algorithmacy/FLAGS.md new file mode 100644 index 00000000..b51a4c0f --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: *Personnel Psychology* + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **No data exist.** The packet says "no data is in hand at the time of writing" (`STUDY.md`), and `PRE_REGISTRATION.md` line 17 still has a placeholder commit date. The paper can be submitted only after waves 1–3 (week 16 falls around mid-December 2026 if fielding started in September). Confirm with the author whether wave 1 opened. +2. **Say "hypotheses fixed before data collection", not "pre-registered".** No external registry record exists; the git history is the only evidence. +3. **Sample size.** The cohort is "expected tens, not hundreds". A *Personnel Psychology* reviewer will question a CFA on that N. The analysis plan treats N as a feasibility constraint. Consider adding a second sample (for example the GauntleTT cohort, which needs its own IRB). +4. **IRB.** Adding the Zhou block lengthens wave 2. A Research Progress & Review Form, or confirmation, is needed before wave 2 (`PRE_REGISTRATION.md`, lines 142–148). Protocol 260511078 was ruled exempt for the survey design. +5. **Inconsistency.** `PRE_REGISTRATION.md` (lines 93–94) predicts for three Zhou subscales and omits Remediating. Fix before fielding wave 2. +6. **Item wording versus the construct.** Several items refer only to "a tool or system" or "reviewers" (for example `acs_ci_2`, `acs_ci_4`, `acs_rt_1`). The construct's distinguishing feature is the counterpart and the system together. Reviewers, and the OT paper's own critique of inventories, will ask whether the items measure algorithmic competency rather than algorithmacy. Consider revising before wave 1 if it has not opened. +7. **Zhou trap:** the scale is Zhou, Lei, Liu, Huang & Hou (2025, *APJHR*). Never write "Zhou et al. 2025" bare. +8. **Fit.** The call asks about the "lived experience" of GenAI. The developer cohort coordinates through peer review and automated checks, so the GenAI framing must be shown in the setting, not asserted. diff --git a/submissions/perspsych_measuring_algorithmacy/OUTLINE.md b/submissions/perspsych_measuring_algorithmacy/OUTLINE.md new file mode 100644 index 00000000..cd1b44ce --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Algorithmacy at Work (*Personnel Psychology*; empirical article, planned at about 11,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Content | Sources | +|---|---|---|---|---| +| 1 | Introduction | 1,200 | The relational experience of GenAI; the measurement gap. | `hindsvonkrogh2024` | +| 2 | Theory | 2,000 | Coordination through an evaluating system; the three facets derived from what the system removes. Answers the OT paper's critique of inventories by deriving items from the structural criterion. | `zhou2025apjhr`, `dogruel2022`, `spitzberg2006`, `sandberg2000` | +| 3 | Hypotheses | 800 | H1a–c (structure, reliability, invariance), H2 (growth), H3 (nomological), H4 (discriminant, latent r < .85). | `PRE_REGISTRATION.md` | +| 4 | Method | 2,000 | Setting, waves (week 0, 4 and 16), sample, measures and analysis (CFA, latent growth or multilevel, FIML, Benjamini–Hochberg). | `STUDY.md`, `analysis_plan.md`, `codebook.md` | +| 5 | Results | 2,500 | [To come after fielding.] | — | +| 6 | Discussion | 2,000 | Contribution, limits (small N, one cohort, self-report) and next studies. | — | +| 7 | Conclusion | 500 | — | — | +| | **Total** | **11,000** | | | diff --git a/submissions/perspsych_measuring_algorithmacy/README.md b/submissions/perspsych_measuring_algorithmacy/README.md new file mode 100644 index 00000000..6764979d --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/README.md @@ -0,0 +1,9 @@ +# *Personnel Psychology*, "The Generative Workplace": Algorithmacy at Work + +[DRAFTED by Claude, 26 Oct 2026, from the ANZAM repost of the call.] + +**Status:** project opened on 26 Oct 2026, with a planned-study abstract, argument and outline drafted. **The submission window is 1–27 Aug 2027.** It needs fielded data (FLAGS 1). Nothing has been submitted. + +**Angle:** a three-facet scale whose items refer to the counterpart and the system together, tested in a three-wave cohort panel. + +**Relation to other arms:** the instrument lives in `org_frontier/survey/cohort_algorithmacy/`. Its conceptual companion is [`../ecis2027_algorithmacy_is/`](../ecis2027_algorithmacy_is/). diff --git a/submissions/perspsych_measuring_algorithmacy/cfp/README.md b/submissions/perspsych_measuring_algorithmacy/cfp/README.md new file mode 100644 index 00000000..8fb47657 --- /dev/null +++ b/submissions/perspsych_measuring_algorithmacy/cfp/README.md @@ -0,0 +1,19 @@ +# CFP brief: *Personnel Psychology*, "The Generative Workplace: The Lived Experience of Generative AI at Work" + +[DRAFTED by Claude, 26 Oct 2026. The Wiley call page returned HTTP 403. Facts come from the ANZAM repost.] + +**Read on 26 Oct 2026:** https://www.anzam.org/2026/09/papers/call-for-papers-special-issue-the-generative-workplace-the-lived-experience-of-generative-ai-at-work/. Blocked: https://onlinelibrary.wiley.com/page/journal/17446570/call-for-papers/generative-workplace. + +**Guest editors:** Louis Tay, Sarah Bankins, Markus Langer. + +## Dates (verbatim, ANZAM) +- "Deadline: August 27 2027" +- "Submission window: 1 August – 27 August 2027" + +## Scope (verbatim, ANZAM) +- "How does using generative AI change the experience of work, workplace relationships, identity, leadership, and organizational culture?" +- "We seek theory-grounded, practice-relevant, empirically rigorous work on the subjective and relational experience of using generative AI at work, and on the consequences of that experience for workers, teams, and organizations." +- "how people write, design, build, decide, persuade, mentor, evaluate, and relate to one another." + +## Unknown +Length, format and article types are in the blocked Wiley page. diff --git a/submissions/sase2027_platform_cooptation/ABSTRACT.md b/submissions/sase2027_platform_cooptation/ABSTRACT.md new file mode 100644 index 00000000..25e1965d --- /dev/null +++ b/submissions/sase2027_platform_cooptation/ABSTRACT.md @@ -0,0 +1,47 @@ +# Abstract: SASE 2027, MC09 "Technopolitics" (alternative: Network J, 16 Dec) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. SASE expects abstracts of about 1,000 words, not counting references. MC09 is due 1 Nov 2026.] + +**Title:** Governed Without a Seat: Platform Co-optation and the Governance Work of the Governed + +## Abstract +Economic sociology classifies governance by the relation it creates between parties. Hierarchies coordinate by authority, markets by price and contract, and networks by reciprocity and reputation. Platforms have strained this classification. One response treats them as a new form. Stark and Vanden Broeck put it in a verb: actors in hierarchies command, in markets contract and in networks collaborate, while on platforms they are co-opted. Vallas and Schor offer another image, the platform as a *permissive potentate* that holds concentrated power while ceding control over much of the work. This paper takes a third position. Platform co-optation is not a new logic of coordination. It is a governance arrangement with a specific problem, and that problem shifts the work of governance onto the people being governed. + +The arrangement has four features. First, it is three-sided. The operator stands between two outside parties, such as a seller and a buyer, a hotel and its guests, or an app developer and its users, and employs neither. Second, evaluation passes between the parties. The operator enrols each side in rating the other, then fuses those ratings with its own ranking rule into a verdict that binds the rated side. Hotel accreditation shows the scale of the change. Anonymous reviews, aggregated by proprietary algorithms into rankings, displaced inspection-based accreditation such as the Michelin Guides and the AA star system (Orlikowski and Scott, 2023). Third, the mediator is opaque and not neutral. It is a party to the exchange it judges, and its rule is neither published nor stable. Fourth, the governed have no seat and a costly exit. Selznick's co-optation absorbed outsiders into the leadership of an organization in order to secure its stability. Platform co-optation reverses the direction: it projects evaluative authority over outsiders who have no part in setting the rule. Leaving is expensive because ratings earned on one platform do not transfer to another. + +Each classical form gives the parties something to govern with. Employees can appeal to authority, contracting parties can bargain over price, and network partners can draw on reciprocity. Platform co-optation offers none of these. In Stark and Vanden Broeck's terms, the operator leverages the parties' agency rather than delegating to them, and accountability becomes distributed, deflected and denied. The result is a governance gap between the two rated parties. Neither can tell how much of a verdict came from the counterpart and how much from the rule. Neither can call the mediator to account in a forum that hears both. Neither can leave without losing standing. + +The governed fill this gap in two ways, and the paper's contribution is to name both and show how they relate. The first is individual competence. Sellers, hosts and service workers learn to read an ambiguous score for its likely cause. They write messages that satisfy both a human counterpart and the evaluator, and they infer rule changes from shifts in their results. This competence is unevenly distributed, which makes the platform a new source of stratification among formally equal participants. It also has a perverse effect: the more skilfully the governed adapt, the less visible the mediator's authority becomes. The second is collective voice. Hirschman showed that easy exit weakens voice. Non-portable ratings make exit costly, which leaves voice as the remaining lever. Forums and worker organisations pool the evidence that no single participant can see, yet, as Johnston and Land-Kazlauskas report, many platform companies have been unwilling to bargain with workers directly. + +Hotels show the arrangement most completely. A hotel and its guests are the two outside parties. Review platforms enrol guests in rating the hotel and rank hotels on an aggregate that the platform computes and does not publish. Scott and Orlikowski found that this user-generated content subverted the older hierarchical accountability in travel, in which complaints moved through internal lines of authority, and redistributed relations of accountability. A hotel cannot see how a guest's review entered its rank. It cannot take a disputed review to a forum where the guest is also heard, and it cannot carry its standing to a competing site: for properties listed on both Airbnb and TripAdvisor, the two sites' ratings correlate only weakly (Zervas et al., 2021). All four features are present, and so is the governance gap. + +The paper builds this argument from published studies of three settings: online marketplaces, hotel rating platforms, and app stores and other complementor ecosystems. It reads each for the four features and for the two responses. Three implications follow for economic sociology. First, the analysis of platforms should treat the relation between the two rated parties as a site of governance, not only the relation between operator and provider. Second, research on digital inequality should measure competence in addressing an evaluator, not only access and skill with devices. Third, regulation that targets transparency alone will miss the problem. A published rule does not give the governed a seat, a shared forum or a portable record. Those three goods are the objects of the technopolitics the mini-conference asks about. They are also what the governed most lack. + +**Keywords:** platforms; co-optation; governance; evaluation; voice; digital economy + +## References (cited in the abstract) +- Hirschman, A. O. (1970). *Exit, voice, and loyalty: Responses to decline in firms, organizations, and states*. Harvard University Press. +- Johnston, H., & Land-Kazlauskas, C. (2018). *Organizing on-demand: Representation, voice, and collective bargaining in the gig economy* (Conditions of Work and Employment Series No. 94). International Labour Office. +- Orlikowski, W. J., & Scott, S. V. (2023). The digital undertow and institutional displacement: A sociomaterial approach. *Organization Theory*, 4(2). https://doi.org/10.1177/26317877231180898 +- Scott, S. V., & Orlikowski, W. J. (2012). Reconfiguring relations of accountability: Materialization of social media in the travel sector. *Accounting, Organizations and Society*, 37(1), 26–40. https://doi.org/10.1016/j.aos.2011.11.005 +- Selznick, P. (1949). *TVA and the grass roots: A study in the sociology of formal organization*. University of California Press. +- Stark, D., & Pais, I. (2020). Algorithmic management in the platform economy. *Sociologica*, 14(3), 47–72. https://doi.org/10.6092/issn.1971-8853/12221 +- Stark, D., & Vanden Broeck, P. (2024). Principles of algorithmic management. *Organization Theory*, 5(2). https://doi.org/10.1177/26317877241257213 +- Vallas, S., & Schor, J. B. (2020). What do platforms do? Understanding the gig economy. *Annual Review of Sociology*, 46, 273–294. https://doi.org/10.1146/annurev-soc-121919-054857 +- Zervas, G., Proserpio, D., & Byers, J. W. (2021). A first look at online reputation on Airbnb, where every stay is above average. *Marketing Letters*, 32(1), 1–16. https://doi.org/10.1007/s11002-020-09546-4 + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Command, contract, collaborate, co-opt | `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 1); see also `dxd_foundation/literature/library/stark2020algorithmic.md` (p. 48) | full text | +| *permissive potentate* | `lima_pdw/literature/cards/vallas2020.md` | full text | +| Three-sided relation | `dxd_foundation/literature/library/stark2020algorithmic.md` (p. 53) | full text | +| Leverages rather than delegates; distributed, deflected, denied | `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) | full text | +| Rankings displaced hotel accreditation | `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3) | full text | +| Selznick's definition | `lima_pdw/literature/cards/selznick1949.md` (p. 13) | extended | +| Ratings do not travel | `hospitality_phygital/library/cards/zervas2021reputation.md` | full text | +| Complementor power asymmetry | `coordinative_sovereignty/literature/library/cards/cutolo2021platform.md` | full | +| Voice where exit is costly | `lima_pdw/literature/cards/hirschman1970.md` | full text | +| Platforms unwilling to bargain | `coordinative_sovereignty/literature/library/cards/johnston2018organizing.md` (p. 23) | full | +| Ratings subvert hierarchical accountability in travel | `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` | full text | +| Hierarchy, market, network (named, not cited) | `lima_pdw/literature/cards/powell1990.md` | secondary summary only (FLAGS 3) | diff --git a/submissions/sase2027_platform_cooptation/ARGUMENT.md b/submissions/sase2027_platform_cooptation/ARGUMENT.md new file mode 100644 index 00000000..df8f2fb1 --- /dev/null +++ b/submissions/sase2027_platform_cooptation/ARGUMENT.md @@ -0,0 +1,19 @@ +# Argument: Platform co-optation as a governance arrangement + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Economic sociology describes governance by the relation it sets up between parties: authority in hierarchies, price and contract in markets, reciprocity and reputation in networks. Platforms are often read as a hybrid of these forms. +- *Sources:* `lima_pdw/literature/cards/powell1990.md` (secondary summary only, see FLAGS); `lima_pdw/literature/cards/vallas2020.md` ("permissive potentates"). + +**P2.** Platforms govern two outside parties through evaluation that passes between them. The operator enrols both in rating and ranking, fuses their judgements with its own opaque rule, and binds each to the result. It does this while holding neither as employee nor agent, and while ratings stay non-portable. +- *Sources:* + - `dxd_foundation/literature/library/stark2020algorithmic.md` (pp. 48, 53, 61) + - `lima_pdw/literature/cards/starkvandenbroeck2024.md` (p. 14) + - `lima_pdw/literature/cards/curchod2020.md` + - `lima_pdw/literature/cards/orlikowskiscott2023.md` (p. 3) + - `hospitality_phygital/library/cards/zervas2021reputation.md` + - `coordinative_sovereignty/literature/library/cards/cutolo2021platform.md` + +**C.** What is new about platform co-optation is not a fourth logic of coordination. It is a governance problem: an interested, opaque mediator holds evaluative authority over parties who have no seat, no shared forum and a costly exit. The arrangement therefore shifts the work of governance onto the governed. They need individual competence to read and address the mediator, and collective voice to contest it. +- *Warrant for the step:* this follows from P2's four features (interest, opacity, no seat, costly exit), set against P1's forms, each of which gives the parties either authority, a price or reciprocity to work with. +- *Lab constraint:* H5 REFRAME (`coordinative_sovereignty/research/research_database.md`, line 23). diff --git a/submissions/sase2027_platform_cooptation/FLAGS.md b/submissions/sase2027_platform_cooptation/FLAGS.md new file mode 100644 index 00000000..84b4033d --- /dev/null +++ b/submissions/sase2027_platform_cooptation/FLAGS.md @@ -0,0 +1,15 @@ +# Flags: SASE 2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The route has to be chosen.** MC09 is due 1 Nov 2026 and Network J is due 16 Dec 2026. "No paper ... may be submitted twice to different networks or mini-conferences." Network J's call (participatory evaluation systems, digital reputations) is the closer fit, and it leaves time. +2. **MC04 is not the algorithmic-management mini-conference.** The 26 Oct search report was wrong on this; corrected in `cfp/README.md`. +3. **Powell (1990) is held only as a secondary summary.** The abstract names the three forms without quoting Powell. Read a clean primary before the full paper. +4. **H5 REFRAME:** the abstract says co-optation is "not a new logic of coordination". This follows the lab's ruling. Stark & Pais are cited for the verb, not for the claim of a fourth form. +5. **Stark & Pais (2020) is not cited in the abstract;** the verb contrast is quoted from Stark & Vanden Broeck (2024, p. 1). If Stark & Pais is cited in the paper, its year is unsettled. The volume says 2020 and the PDF header says January 2021. Cite 2020 with the volume. +6. **Johnston & Land-Kazlauskas (2018)**: the PDF front matter says 2019; the card keeps 2018 pending an imprint check. The abstract no longer claims specific campaigns for rule notice or joint hearings, because none is carded. +7. **App stores and complementors** rest on the Cutolo & Kenney card (abstract quote, verified full). No app-store case study is carded yet. +8. **The hotel example** uses Orlikowski & Scott (2023), which is pitched at industry level. It supports displacement of accreditation, not anything about individual workers. +9. **Kornberger, Pflueger & Mouritsen (2017)** is the earlier statement of the platform as coordinated by ratings. It is held only as metadata (`coordinative_sovereignty/literature/library/cards/kornberger2017evaluative.md`). Read it before the full paper, because reviewers will expect it. +10. **Hirschman** is about members who already hold membership. Platform exit that forfeits non-portable ratings is the author's extension, not his model (card caution). +11. **Powell** is named in the abstract without a citation, because only a secondary summary is held. diff --git a/submissions/sase2027_platform_cooptation/OUTLINE.md b/submissions/sase2027_platform_cooptation/OUTLINE.md new file mode 100644 index 00000000..74af359d --- /dev/null +++ b/submissions/sase2027_platform_cooptation/OUTLINE.md @@ -0,0 +1,14 @@ +# Outline: Governed Without a Seat (SASE 2027 full paper) + +[DRAFTED by Claude, 26 Oct 2026. SASE recommends but does not require a full paper. This outline sizes one at about 8,000 words, which is also a journal-length draft for *Socio-Economic Review* or a similar journal.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 800 | Covers the two readings (new form, hybrid) and states the third position: a governance arrangement that shifts governance work onto the governed. | `stark2020algorithmic`, `vallas2020` | +| 2 | What each form gives the parties to govern with | 1,000 | Authority, price and reciprocity. A table sets form, coordinating relation, and the parties' lever. | `powell1990` (needs a clean primary), `selznick1949` | +| 3 | Four features of platform co-optation | 1,800 | Three sides; evaluation passing between the parties; an interested, opaque mediator; no seat and costly exit. One published case for each. | `stark2020algorithmic`, `starkvandenbroeck2024`, `curchod2020`, `orlikowskiscott2023`, `zervas2021reputation`, `cutolo2021platform` | +| 4 | The governance gap between the rated parties | 1,200 | No split verdict, no shared forum and no portable standing. | `rahman2021` (theory only), `scottorlikowski2012accountability`, `bovens2007` | +| 5 | Individual competence as governance work | 1,200 | The three practices, stratification among formal equals, and the masking effect. | `devito2021`, `jhaver2018algorithmic`, `hargittai2002`, `elish2019` | +| 6 | Collective voice | 1,100 | Voice where exit is costly; pooled evidence; demands for notice and joint hearing. | `hirschman1970`, `johnston2018organizing`, `rahman2024taming` (abstract) | +| 7 | Implications | 900 | For platform analysis, digital-inequality research and transparency regulation. | `ananny2018` (abstract), `citronpasquale2014scored` (abstract) | +| | **Total** | **8,000** | | | diff --git a/submissions/sase2027_platform_cooptation/README.md b/submissions/sase2027_platform_cooptation/README.md new file mode 100644 index 00000000..0ad5c439 --- /dev/null +++ b/submissions/sase2027_platform_cooptation/README.md @@ -0,0 +1,29 @@ +# SASE 2027: Governed Without a Seat + +[DRAFTED by Claude, 26 Oct 2026, from the organizers' pages. The angle and route are proposals for the author to rule on.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **The MC09 abstract (about 1,000 words) is due 1 Nov 2026. The alternative, Network J, is due 16 Dec 2026.** Nothing has been submitted. + +**The call:** 39th SASE Annual Conference, Glasgow, 30 June – 3 July 2027. Digest: [`cfp/README.md`](cfp/README.md). + +| Milestone | Date (from the call) | +|---|---| +| Mini-conference abstract (MC09 Technopolitics) | **1 Nov 2026** | +| Network abstract (Network J, Digital Economy) | 16 Dec 2026 | +| Conference | 30 June – 3 July 2027 | + +**Angle:** platform co-optation is a governance arrangement, not a fourth logic. It shifts governance work onto the governed through individual competence and collective voice. + +**Relation to other arms:** +- Builds on [`../coordinative_sovereignty/`](../coordinative_sovereignty/) (H5 REFRAME) and [`../algorithmacy_ot/`](../algorithmacy_ot/). +- Its sibling [`../ilpc2027_cooptation_control/`](../ilpc2027_cooptation_control/) argues control, not governance. + +| File | What it is | +|---|---| +| `ABSTRACT.md` | Title, abstract (824 words; SASE expects about 1,000), references and citation check | +| `ARGUMENT.md` | P1, P2, C | +| `OUTLINE.md` | 7-section outline, about 8,000 words | +| `FLAGS.md` | Route choice and evidence cautions | +| `cfp/` | Digest; `source/` is gitignored | + +**Guardrails:** Claude does not submit. The author's read-aloud is the last gate. diff --git a/submissions/sase2027_platform_cooptation/cfp/README.md b/submissions/sase2027_platform_cooptation/cfp/README.md new file mode 100644 index 00000000..c3dcf8b4 --- /dev/null +++ b/submissions/sase2027_platform_cooptation/cfp/README.md @@ -0,0 +1,26 @@ +# CFP brief: SASE 2027 + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the organizers' pages, not a substitute for them.] + +**Read on 26 Oct 2026:** +- https://sase.org/events/2027-glasgow/ (only the first half of the page was verbatim: MC01–MC10. The rest of the extraction was summarised.) +- https://sase.org/conference-submission-guidelines/ +- https://sase.org/deadlines/ +- https://sase.org/networks/j-digital-economy/ + +**Event:** 39th SASE Annual Conference, "Rebuilding Foundations and Hope: Creative Reconstruction in a Moment of Disruption". University of Strathclyde, Glasgow, 30 June – 3 July 2027; virtual events 16–18 June 2027. + +## Dates (verbatim) +- "The final deadline for submissions to mini-conferences is the 1st of November, 2026. This is different to previous years and earlier than submissions to networks (16th of December, 2026)." +- No timezone is stated. + +## Format (verbatim) +- "To submit an individual presentation, you must include an abstract. Abstracts are expected to be 1,000 words, not including references." +- Mini-conferences recommend, but do not require, a full paper before the conference. +- "No paper or panel may be submitted twice to different networks or mini-conferences." "no author may present more than two papers". +- Portal: Oxford Abstracts. + +## Correction to the 26 Oct search +MC04 is "Working time reduction", not algorithmic management. The fitting homes are: +- **MC09 "Technopolitics: Power, Competition and Contestation over Digital Infrastructures"** (Anwar, Rolf, Thelen, Cross, Lehuedé, Taylor, O'Reilly). Its call says: "Digital infrastructures (broadly defined) have become unavoidable and powerful intermediaries through which our socio-economic, political and ecological relations are organised." Its indicative panel is "Labour, class and exploitation in algorithmic capitalism and AI". **Due 1 Nov 2026.** +- **Network J, Digital Economy** (Beuscart, Ergen, Jourdain, Lehdonvirta, Mellet, Naulin, Vargha, Moreno). It covers "intermediaries and business models such as the platform economy, coordination instruments such as participatory evaluation systems, digital reputations". **Due 16 Dec 2026.** diff --git a/submissions/sij_service_triad_ethics/ABSTRACT.md b/submissions/sij_service_triad_ethics/ABSTRACT.md new file mode 100644 index 00000000..5d5b5d74 --- /dev/null +++ b/submissions/sij_service_triad_ethics/ABSTRACT.md @@ -0,0 +1,19 @@ +# Abstract: *The Service Industries Journal* special issue + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. Due 30 June 2027.] + +**Title:** Where Accountability Breaks in the Service Triad, and to Whom It Can Be Returned + +## Abstract +Service work has always involved three parties: the worker, the customer and the firm. On booking and home-sharing platforms, and increasingly inside hotels, an evaluating system now occupies the third corner. It turns guests' judgements into verdicts on staff, hosts and properties, and it sets the rules under which guests and hosts meet. This paper asks where accountability breaks in that arrangement and which mechanisms can restore it. Drawing on published studies of hotel reviews, home-sharing hosts, discrimination in guest acceptance and portability of reputation, we identify four break points. Score aggregation hides how much of a verdict each party produced. Access rules set by the intermediary sort guests before any service begins. Ratings do not travel between platforms, so a poor verdict cannot be escaped. Oversight assigned to frontline staff who cannot override the system leaves them absorbing responsibility without control. We then compare four restoring mechanisms by the party to which each returns the verdict. Appeals return it to the evaluated worker or property. Reciprocal reveal returns it to both parties at once. Managerial adjustment returns it to the firm. Guest-visible attribution returns it to the customer. The resulting framework tells service managers and regulators which mechanism fits which break. It also shows why transparency alone, which returns the verdict to no one, does not restore accountability in the service triad. + +**Keywords:** service triad; algorithmic accountability; hospitality platforms; opacity; human oversight + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Hotel reviews subvert hierarchical accountability | `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` | full text | +| Hosts' double negotiation | `lima_pdw/literature/cards/jhaver2018.md` | full text | +| Discrimination in guest acceptance; harm in the intermediary's design | `hospitality_phygital/library/cards/edelman2017discrimination.md` | full text | +| Reputation does not travel | `hospitality_phygital/library/cards/zervas2021reputation.md` | full text | +| Oversight without authority to override | `hospitality_phygital/library/cards/fink2025oversight.md` | full text (working paper) | diff --git a/submissions/sij_service_triad_ethics/ARGUMENT.md b/submissions/sij_service_triad_ethics/ARGUMENT.md new file mode 100644 index 00000000..ea9b26fe --- /dev/null +++ b/submissions/sij_service_triad_ethics/ARGUMENT.md @@ -0,0 +1,18 @@ +# Argument: where accountability breaks in the service triad + +[DRAFTED by Claude, 26 Oct 2026. Angle sharpened against the OT paper (see FLAGS 2).] + +**P1.** Service work has long been a triangle of worker, customer and firm. Platforms make the third corner an evaluating system that turns customers' judgements into verdicts on workers and properties. The evaluated side often cannot see how the verdict was formed. +- *Sources:* + - `hospitality_phygital/library/cards/scottorlikowski2012accountability.md` + - `lima_pdw/literature/cards/jhaver2018.md` (hosts' double negotiation) + - `dxd_foundation/literature/library/cruz2023digital.md` (p. 2: platform–client–worker as a triangular service relation) + +**P2.** Accountability breaks at identifiable points in that triad. Score aggregation hides each party's contribution. Access rules set by the intermediary sort guests before any service begins. Ratings do not travel between platforms. Oversight given to staff who cannot override the system makes them absorb responsibility without control. +- *Sources:* + - `lima_pdw/literature/cards/rahman2021.md` (p. 976, paraphrase) + - `hospitality_phygital/library/cards/edelman2017discrimination.md` (the harm located in the intermediary's design decision) + - `hospitality_phygital/library/cards/zervas2021reputation.md` + - `hospitality_phygital/library/cards/fink2025oversight.md` (liability sponge; working paper) + +**C.** Mechanisms that restore accountability differ in which party they return the verdict to. Appeals return it to the evaluated worker or property. Reciprocal reveal returns it to both parties at once. Managerial adjustment returns it to the firm. Guest-visible attribution returns it to the customer. Service firms and regulators should choose mechanisms by where the break occurs and to whom the verdict should return. diff --git a/submissions/sij_service_triad_ethics/FLAGS.md b/submissions/sij_service_triad_ethics/FLAGS.md new file mode 100644 index 00000000..0bc73cb9 --- /dev/null +++ b/submissions/sij_service_triad_ethics/FLAGS.md @@ -0,0 +1,12 @@ +# Flags: *SIJ* + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The word limit is unknown** because the author instructions were blocked. +2. **Overlap with the OT paper is medium.** The OT paper already uses the service-triangle lineage (Fuller & Smith, Lopez), the hotel rankings and Airbnb's simultaneous reveal. This paper's contribution is the break-point × mechanism framework. Use new cases (Edelman, Zervas, Fink) and do not reuse the OT paper's wording. +3. **Overlap with `../isj_future_of_work_design/`:** that paper owns artifact design properties. This one stays with governance and organizational mechanisms, and must not propose interface features. +4. **"Reciprocal reveal"** refers to Airbnb's simultaneous review reveal (Fradkin et al., 2021), which has no card. Card it before writing §4. +5. **Fink (2025) is an SSRN working paper.** Check for a journal version. The card advises pairing it with Sterz (2024), which has a card in the hospitality library (not read here). +6. **Edelman et al.** concentrate the effect among hosts with no prior African-American guest, and decline to separate race from socioeconomic status. +7. **The hotel-internal systems ("increasingly inside hotels")** have no specific card. Soften or source before submission. +8. **"Managerial adjustment" as a mechanism** rests on the OT paper's Lin et al. (2026) citation, which has no card here. diff --git a/submissions/sij_service_triad_ethics/OUTLINE.md b/submissions/sij_service_triad_ethics/OUTLINE.md new file mode 100644 index 00000000..93b2a0d5 --- /dev/null +++ b/submissions/sij_service_triad_ethics/OUTLINE.md @@ -0,0 +1,13 @@ +# Outline: the service triad (*SIJ*; length to confirm, planned at about 8,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Introduction | 800 | The service triangle and its new third corner. | `scottorlikowski2012accountability`, `cruz2023digital` | +| 2 | From service triangle to evaluating intermediary | 1,300 | Service-work lineage; platforms and hotel systems. | `cruz2023digital`, `vandoorn2023` (abstract; CAW rival), `munasinghe2022` (abstract) | +| 3 | Four break points | 2,200 | Aggregation, access rules, non-portability and oversight without authority, each with a published hospitality case. | `rahman2021` (theory only), `edelman2017discrimination`, `zervas2021reputation`, `teubner2019reputation` (abstract), `fink2025oversight`, `jhaver2018` | +| 4 | Four restoring mechanisms | 1,800 | Appeal, reciprocal reveal, managerial adjustment and guest-visible attribution, by recipient of the verdict. | `alfrink2023contestable`, `bovens2007` | +| 5 | Framework and propositions | 1,100 | A matrix of break by mechanism. Why transparency alone returns the verdict to no one. | — | +| 6 | Implications and limits | 800 | For managers and regulators (EU AI Act Art. 14). Conceptual and based on secondary cases. | `fink2025oversight` | +| | **Total** | **8,000** | | | diff --git a/submissions/sij_service_triad_ethics/README.md b/submissions/sij_service_triad_ethics/README.md new file mode 100644 index 00000000..f5bb07cc --- /dev/null +++ b/submissions/sij_service_triad_ethics/README.md @@ -0,0 +1,9 @@ +# *The Service Industries Journal*: Where Accountability Breaks in the Service Triad + +[DRAFTED by Claude, 26 Oct 2026, from the call page.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **Due 30 June 2027.** Nothing has been submitted. + +**Angle:** four break points of accountability in the worker–guest–platform triad, and four restoring mechanisms classified by the party to which each returns the verdict. + +**Relation to other arms:** draws on [`../hospitality_phygital/`](../hospitality_phygital/) (162 cards). It is distinct from [`../isj_future_of_work_design/`](../isj_future_of_work_design/) and [`../jms_organizing_truth_arbiter/`](../jms_organizing_truth_arbiter/). diff --git a/submissions/sij_service_triad_ethics/cfp/README.md b/submissions/sij_service_triad_ethics/cfp/README.md new file mode 100644 index 00000000..9673d7ff --- /dev/null +++ b/submissions/sij_service_triad_ethics/cfp/README.md @@ -0,0 +1,21 @@ +# CFP brief: *The Service Industries Journal*, "Ethical implications of artificial intelligence (AI) and automation in service industries: Addressing algorithmic bias, opacity and unclear accountability mechanisms" + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the call page, not a substitute for it.] + +**Read on 26 Oct 2026:** https://think.taylorandfrancis.com/special_issues/ethical-implications-of-artificial-intelligence-ai-and-automation-in-service-industries-addressing-algorithmic-bias-opacity-and-unclear-accountability-mechanisms/. The author instructions (tandfonline, journal code FSIJ) returned HTTP 403. + +**Guest editors:** Mark Anthony Camilleri (Malta, leading), Serena Strazzullo (Naples Federico II), Raffaele Campo (Bari), Alhamzah F. Abbas (UTM), Pasquale Del Vecchio (LUM), Mark S. Rosenbaum (The Citadel). + +## Dates (verbatim) +- "Manuscript deadline: 30 June 2027" ("Full paper submission deadline: 30th June 2027"). No timezone is stated. + +## Format +- "Manuscripts should be prepared according to The Service Industries Journal's author guidelines ... During submission, authors should select the special issue title". Double-blind review. +- Word limit: not stated (the instructions were blocked). +- "Both conceptual, theoretical and empirical contributions are welcome, including qualitative, quantitative, mixed-methods, experimental, design science as well as critical and/or reflexive approaches." + +## Lines nearest this paper (verbatim) +- "These systems often operate as "black boxes," leaving workers uncertain about how ratings, penalties or income calculations are determined." +- "Unclear accountability structures may obscure responsibility for service failures or might facilitate unintended harmful outcomes" +- "Algorithmic opacity makes it difficult for customers to understand, question or contest adverse decisions." +- Topics: "Accountability and responsibility in human–AI service interactions"; "Human oversight and hybrid human–AI service models"; "Transparency and explainability in automated service decisions". diff --git a/submissions/spm2027_lived_arbiter/ABSTRACT.md b/submissions/spm2027_lived_arbiter/ABSTRACT.md new file mode 100644 index 00000000..979ffa0b --- /dev/null +++ b/submissions/spm2027_lived_arbiter/ABSTRACT.md @@ -0,0 +1,17 @@ +# Abstract: SPM2027, individual paper (Axis 1) + +[DRAFTED by Claude, 26 Oct 2026. For the author's read-aloud. 150–250 words plus a 50–100-word biography. Due 15 Dec 2026.] + +**Title:** The Lived Arbiter: Speaking to Another Person Before an Invisible Judge + +## Abstract +Accounts of bureaucratic opacity usually describe one person facing an institution. On digital platforms, opacity is lived in a three-person situation. A host speaks to a guest and a seller writes to a buyer, and each is heard by a ranking algorithm that judges them both. This paper takes its framework from Simmel's sociology of the third. The mediator leaves the decision with the parties, while the arbitrator receives it from them. A third may also profit from the parties' division. The platform algorithm is an arbitrator that neither party chose and neither can see. Drawing on published interview studies of Airbnb hosts and of reviewers on a filtered review site, the paper describes how this arbiter is lived. Hosts report a "double negotiation", with guests and with partly transparent algorithms, and an anxiety that attaches to the unseen judge rather than to the guest. Reviewers who learn of a hidden filter begin to write for it as an audience. The paper argues that platform institutionality is experienced as a doubled address: every word to the other person is also a submission to the arbiter. + +**Biography (50–100 words):** [AUTHOR TO WRITE. Claude does not draft the biography.] + +## Citation check +| Claim | Source card | Depth | +|---|---|---| +| Mediator versus arbitrator; the third who profits | `dxd_foundation/literature/library/simmel1950sociology.md` (pp. 148, 151, 154) | full text, pp. 135–169 | +| Hosts' "double negotiation" with guests and algorithms | `dxd_foundation/literature/library/jhaver2018algorithmic.md` (abstract) | full text | +| Writing for the filter as audience | `dxd_foundation/literature/library/eslami2019user.md` (p. 11) | full text | diff --git a/submissions/spm2027_lived_arbiter/ARGUMENT.md b/submissions/spm2027_lived_arbiter/ARGUMENT.md new file mode 100644 index 00000000..e2df756c --- /dev/null +++ b/submissions/spm2027_lived_arbiter/ARGUMENT.md @@ -0,0 +1,11 @@ +# Argument: the lived arbiter + +[DRAFTED by Claude, 26 Oct 2026. For the author's ruling.] + +**P1.** Simmel distinguishes the mediator, who leaves the decision with the parties, from the arbitrator, to whom the parties hand it over. A third can also benefit from the parties' division, or produce it. +- *Sources:* `dxd_foundation/literature/library/simmel1950sociology.md` (pp. 148, 151, 154, 162); `triadic_reduction/library/cards/simmel1902number2.md` (pp. 169, 172). + +**P2.** A platform's evaluating algorithm is lived as an arbitrator that nobody chose and nobody can see. Hosts describe a double negotiation, with guests and with only partly transparent algorithms, and the anxiety that follows. Users who learn that a review filter exists begin to write for it as an audience. +- *Sources:* `dxd_foundation/literature/library/jhaver2018algorithmic.md` (abstract); `dxd_foundation/literature/library/eslami2019user.md` (pp. 11–12); `lima_pdw/literature/cards/burrell2016.md` (abstract depth). + +**C.** Institutional opacity on platforms is lived as a three-person situation, not as a relation between one person and an institution. The other person is present, but only through a judge whose criteria are withheld. A phenomenology of platform institutions must describe this doubled address: speaking to another person while being heard by an arbiter. diff --git a/submissions/spm2027_lived_arbiter/FLAGS.md b/submissions/spm2027_lived_arbiter/FLAGS.md new file mode 100644 index 00000000..24ec5384 --- /dev/null +++ b/submissions/spm2027_lived_arbiter/FLAGS.md @@ -0,0 +1,10 @@ +# Flags: SPM2027 + +[DRAFTED by Claude, 26 Oct 2026. Cautions kept out of the prose.] + +1. **The biography is the author's to write.** +2. **Postphenomenology (Ihde, Verbeek) has no card.** An SPM audience will expect it. Card it before the talk. +3. **Simmel's arbitrator is chosen** by the parties. The abstract marks the platform difference ("neither party chose"). "Dyad" and "triad" are Wolff's words, not Simmel's. +4. **Jhaver et al.:** two authors worked at Airbnb, N = 15, and the paper has no page numbers. "Anxiety that attaches to the unseen judge rather than to the guest" is the author's reading of their algorithmic anxiety finding, not their sentence. +5. **Eslami et al. (2019):** only 3 of 15 interviewees knew of the filter, and "writing for it" is a reported intention. +6. **Fees:** $150 to $250, plus Mexico City travel or a virtual presentation. diff --git a/submissions/spm2027_lived_arbiter/OUTLINE.md b/submissions/spm2027_lived_arbiter/OUTLINE.md new file mode 100644 index 00000000..7f33177b --- /dev/null +++ b/submissions/spm2027_lived_arbiter/OUTLINE.md @@ -0,0 +1,12 @@ +# Outline: SPM2027 talk (about 20 minutes) and a *Glimpse* paper (about 6,000 words) + +[DRAFTED by Claude, 26 Oct 2026.] + +| § | Section | Words (paper) | Mechanism | Cards | +|---|---|---|---|---| +| 1 | Opacity lived alone versus lived in three | 900 | Contrasts the single-person bureaucracy encounter with the platform situation. | `bovens2002street` (abstract), `burrell2016` (abstract) | +| 2 | Simmel's third | 1,300 | Mediator, arbitrator, *tertius gaudens* and *divide et impera*. | `simmel1950sociology`, `simmel1902number2` | +| 3 | The arbiter nobody chose | 1,200 | Platform evaluation as an unchosen, invisible arbitrator. | `stark2020algorithmic` (pp. 53–54), `starkvandenbroeck2024` | +| 4 | The doubled address | 1,500 | Hosts' double negotiation and reviewers writing for the filter. Anxiety directed at the unseen judge. | `jhaver2018algorithmic`, `eslami2019user`, `bucher2017` (abstract) | +| 5 | Toward a phenomenology of platform institutions | 1,100 | What a three-place description adds to accounts of institutional opacity. | — | +| | **Total** | **6,000** | | | diff --git a/submissions/spm2027_lived_arbiter/README.md b/submissions/spm2027_lived_arbiter/README.md new file mode 100644 index 00000000..f2a2c27a --- /dev/null +++ b/submissions/spm2027_lived_arbiter/README.md @@ -0,0 +1,11 @@ +# SPM2027: The Lived Arbiter + +[DRAFTED by Claude, 26 Oct 2026, from the EasyChair call.] + +**Status:** project opened on 26 Oct 2026, with an abstract, argument and outline drafted. **The abstract (150–250 words plus a biography) is due 15 Dec 2026.** Nothing has been submitted. + +**The call:** Society for Phenomenology and Media, Mexico City, 16–20 Mar 2027. Digest: [`cfp/README.md`](cfp/README.md). + +**Angle:** platform opacity is lived as a doubled address before an arbiter nobody chose (Simmel's third). + +**Relation to other arms:** draws on [`../triadic_reduction/`](../triadic_reduction/) (Simmel cards). It is distinct from [`../cepe2027_third_party_accountability/`](../cepe2027_third_party_accountability/), which argues accountability's logical form. diff --git a/submissions/spm2027_lived_arbiter/cfp/README.md b/submissions/spm2027_lived_arbiter/cfp/README.md new file mode 100644 index 00000000..5f28c4df --- /dev/null +++ b/submissions/spm2027_lived_arbiter/cfp/README.md @@ -0,0 +1,23 @@ +# CFP brief: Society for Phenomenology and Media, 28th Annual Conference (SPM2027) + +[DRAFTED by Claude, 26 Oct 2026. A working digest of the EasyChair call, not a substitute for it.] + +**Read on 26 Oct 2026:** https://easychair.org/cfp/SPM2027. The society's linked page held only accommodation details. + +**Event:** "SPM2027: The Phenomenology of Institutions. Media, Structures, and Lived Experience." Universidad Panamericana, Mexico City, 16–20 March 2027. Hybrid options are available. Program chair: Mariano Navarro. + +## Dates (verbatim) +- "Deadline for abstracts: December 15, 2026". No timezone is stated. +- "Acceptance rollout: January 15, 2027" +- "Deadline for conference registration (through PayPal): February 1, 2027" + +## Format (verbatim) +- "Individual Paper Proposals: 150-250 word abstract including title, key methodology/framework, and a short author biography (50-100 words)." +- "A maximum of two submissions per author will be accepted". +- Submit through EasyChair (conf=spm20270). Fees: $250 regular; $150 for graduate students, postdocs, contingent faculty and virtual presenters. +- "Selected papers will be invited to expand their work for peer-reviewed publication in SPM's journal, Glimpse." + +## Lines nearest this paper (verbatim) +- Axis 1, "Mediated Institutionality and Platform Governance": "Digital platforms, social networks, and AI systems as emerging institutional structures." +- "Algorithmic governance, automated decision-making, and the lived experience of bureaucratic opacity." +- Axis 2: "The affectivity of institutions: institutional trust, anxiety, alienation, and sense of belonging."