Events as a Stress Test

Note: Do not confuse with Illinois HB3773 / IHRA AI employment amendments effective 2026, already separately tracked.

Note: Not generative AI; strong algorithmic entry-channel signal.

Note:

Note: Non-AI but high-value baseline for workforce data governance and employee-surveillance harms.

Note: Vendor rollback, not court finding. Keep as B-level event, not proof of illegal discrimination.

Note: Positive capability signal; pair with later IP/security/provenance cases.

Note:

Note: Secondary source used for English-language summary; verify against Spanish BOE if used legally.

Note: Technical standard, not an adoption event. Pair with platform adoption events.

Note:

Note:

Note: Allegations unresolved/partly narrowed; provenance signal.

Note: Trust/provenance/quality signal, not labor displacement.

Note: This is a policy-pivot signal, not a detector event. Useful in the broader provenance-collapse response taxonomy.

Note:

Note: Use 2023 filing as event; later 2025 UK ruling should be a separate follow-up if needed.

Note: Quality/provenance signal; layoffs had multiple causes.

Note: Governance/control baseline, not harm signal.

Note: Pair with SIG_2023_OPENAI_AI_CLASSIFIER_RETIRED. This is not a separate contradiction; it shows launch-time caveating followed by retirement after classroom uptake.

Note: Useful counterexample: not all institutions responded with bans or detectors; some shifted to attribution and AI literacy.

Note: Existing v0.2 had AI/layoff signals; this adds explicit manager→IC track-switch evidence.

Note: Research signal; not tied to one institutional deployment. Useful as mechanism evidence for the detector-collapse narrative.

Note: Use with care: PxDx described as algorithmic automation; Cigna disputed framing.

Note: Platform access / regulatory volatility signal.

Note: Pair with Vanderbilt disablement and GPT-detector bias signals; do not treat detector outputs as ground truth.

Note:

Note: Anticipated displacement; not realized one-for-one replacement.

Note: Month-level date; report is dated May 2023.

Note: Market signal, not worker-level displacement.

Note: Code as medium-high because reporting is strong but not a formal court/regulatory record. The key signal is not Turnitin-specific; it is provenance panic and unreliable detection practice.

Note: Do not code all 55,000 as AI replacement.

Note: Complements the existing EEOC/DOJ ADA AI guidance signal.

Note: Vendor evidence; useful but should be treated with vendor-bias caveat. Strong for chronology and stated limitations.

Note: Code as restructuring/role-boundary signal; avoid claiming direct replacement unless source-specific.

Note: Seed signal-preservation metric: sanctions, client notice, bar referral.

Note: Governance signal; does not prove audit effectiveness.

Note:

Note: Strong A-level research signal; use as caution against naive AI-detector-based credential decisions.

Note: Retirement date from OpenAI note; launch date was Jan. 31, 2023.

Note: Voluntary commitments, not binding law.

Note:

Note: Useful as a named institutional decision, not a universal higher-education pattern.

Note: Use as guidance/policy source, not as measured outcomes.

Note: High-confidence settlement; not prevalence estimate.

Note: Confidence lower because available English source is secondary; should be upgraded only with French primary union/company documents.

Note: Allegations unresolved; classify as litigation signal.

Note: Useful collective-agency and D08 signal-preservation precedent.

Note: Version with 2025 revocation; not stable current rule.

Note: Signal preservation for creative-project markets.

Note: Allegations pending/contested; code as litigation signal.

Note: Primary legal complaint should be pinned before litigation use.

Note: Code as surveillance/governance signal, not pure hiring case.

Note: Litigation unresolved; provenance/value signal.

Note: Event date is decision date; public reporting followed in Jan 2024.

Note: Good D10 test case, but not return-panel evidence.

Note: Corporate partnership; use as adoption signal, not outcome evaluation.

Note:

Note: Professional workflow responsibility, not direct displacement.

Note:

Note: Corporate self-report; outsourcing status matters.

Note: Use as regulatory/enforcement signal. Primary official legal acts should be pinned in a later legal appendix.

Note:

Note: Disclosure only for realistic altered/synthetic content; not blanket AI disclosure.

Note:

Note:

Note: Pair with July 2024 label adjustment if tracking false-positive effects.

Note: Unresolved litigation; allegations not findings.

Note: Counts are detector outputs, not verified misconduct. Use as adoption/load signal, not ground truth about cheating prevalence.

Note: Platform adoption event; coverage depends on upstream tools embedding credentials.

Note: Technical release; real-world reliability and adversarial robustness should be tracked separately.

Note: Track amendments/effective-date changes separately.

Note: Contested allegation; note OpenAI denial and voice pause.

Note: Event date approximates mid-2024 collapse/CEO removal; founder arrest occurred in Nov. 2024. Keep as medium_high because some project-timeline details are reconstructed from reporting.

Note: Litigation signal; allegations remain contested.

Note: Not an AI-detector evaluation only; evaluates exam-system infiltration and human-marker detection. Complements detector false-positive cases with false-negative/undetected-use evidence.

Note: Date inferred from Turnitin guide wording that pre-July-8-2024 reports may still show sub-20 scores. Treat as product-policy change, not independent validation.

Note: Skill-mix reallocation; avoid one-for-one AI replacement claim.

Note: Separate corporate transaction (A/B) from staff-count claims (B/C depending on source).

Note: Treat as classification/rights signal, not as evidence of worker preference.

Note: Project-based labor makes it useful for D10 as well as D08.

Note: Official ABA opinion URL should be pinned if available; current source is ABA Business Law summary.

Note: Use published date July 31, 2024 from Copyright Office NewsNet.

Note: Implementation timing must be versioned.

Note: Single-source reporting; code as medium confidence and not as confirmed product policy. Complements Sadasivan/evasion concerns and C2PA/SynthID adoption signals.

Note: Use as introduced-bill signal, not enacted law.

Note: Official memo; code as policy intent, not realized individual outcomes.

Note: Country/system-level only; do not use as individual agency score.

Note:

Note: Bill cites mental-health concerns as primary rationale; AI/cheating concerns secondary in legal text but central in implementation discourse.

Note: Policy signal; not evidence of individual outcomes.

Note: Product/tooling event, not broad adoption evidence.

Note:

Note:

Note:

Note:

Note:

Note: Industry self-report via Pollfish; sample 948 leaders at companies >21 employees. Selection bias toward AI-aware respondents likely.

Note: Use as adjacent-domain scoring benchmark; do not over-map to HR.

Note:

Note: Duplicate-adjacent to v0.2 Council-adoption signal, but this is a separate implementation-clock event.

Note: Data-release anchor; not an event of harm/displacement.

Note: Investigation, not official dataset. Use for lived-experience and institutional-process stress-test, not prevalence estimate.

Note:

Note: Internal all-hands reporting; classify B, not A.

Note: C-level evidence; company statement and worker-level attrition would improve confidence.

Note: Not explicitly AI-driven; keep confidence medium-high for event, low for AI attribution.

Note: Do not oversimplify as 'AI art never protected'; the report preserves human-authorship pathways.

Note: Corporate-policy signal; worker-experience effects need separate data.

Note: Not an AI-causality event. Keep as quiet-restructuring / measurement-visibility signal.

Note: Code as AI-era restructuring, not direct automation proof.

Note: Platform-owned usage data; valuable but not representative of all labor-market activity.

Note:

Note: Strong A-level court ruling; legal context is professional/legal research, not all AI training.

Note: Use as organizational/employee-response signal, not direct attrition data.

Note: Use as labor-market signal; not a panel and not directly comparable to O*NET/PIAAC.

Note: Corporate statement; verify against actual job postings before using as hard hiring data.

Note: Vendor product signal; use as institutional response category, not proof of efficacy.

Note: Keep a separate event for the 2025-09-01 effective date if timeline testing needs implementation clocks.

Note: Policy-failure event; useful for regulation-diffusion analysis.

Note: Selection bias: likely over-represents politically active/high-skill emigrants. Still stronger than official press claims.

Note: Industry survey; HR self-report; sample frame favors AI-aware HR. The 90% spam-rise figure is a perception measure, not a parsed application count.

Note: Reported internal memo; not a public labor-law rule.

Note: Job postings are demand signals, not realized hires.

Note: Industry survey self-report. Whether managers actually reject AI-only resumes at the rates they claim is not separately validated. The 'personalization' criterion is itself unstable — generative tools can also produce personalized output.

Note: Memo-reported; not independent attrition/outcome data.

Note: Use paired with 2023 IBM hiring-pause signal in v0.2; do not encode viral 8,000 rehired claim as verified.

Note: Company narrative and media framing vary; code as walkback/rehiring signal, not proof of exact displacement count.

Note: Bell-to-bell means no use during class, lunch, or passing periods. Local districts decide storage method.

Note: Pre-publication version; track final version separately if released.

Note: Separate AI-platform competition from search-distribution shocks.

Note: AI attribution indirect; classify event high, AI-causality medium.

Note: Official results not yet published in the checked sources; event is fielding/data-infrastructure, not results.

Note: Industry sales data from Roaring Spring; not all universities surveyed. Trend is real but counter-current persists — some faculty argue blue books cannot substitute for take-home research papers.

Note: Outcome evidence not yet available; classify as policy/curriculum signal.

Note: Not clear whether restrictions were voluntary platform policy or state-directed. Code as platform/exam-integrity signal with medium-high confidence.

Note: Survey/investigation data; use as trend signal, not fully harmonized official statistics.

Note: Do not overstate employment-specificity; TRAIGA is broader than employment.

Note: Use carefully: this is a split ruling, not blanket permission for all AI training uses.

Note: Do not generalize beyond the evidence record; court noted plaintiffs' market-harm theory could matter in a better-developed case.

Note: Effective date reported by practitioner sources as Oct 1, 2025.

Note: Use as restructuring/capex signal, not proof of direct AI replacement.

Note: Task/context-specific; not a general anti-AI productivity result.

Note:

Note: Early-stage litigation. Allegations are unproven; keep confidence medium.

Note: Working paper; strong data, still subject to ongoing academic debate.

Note: C-level corporate clarification; code as narrative/walkback, not independent employment data.

Note: Pre-print (Aug 2025); peer-reviewed publication pending in M&SOM. 23-60% range from simulation, not field data. LLM-vs-LLM bias is weaker than LLM-vs-human bias.

Note: Code as CEO disclosure; not a public WARN/filing count.

Note: Corrected from uploaded candidate’s -991,000 to verified -911,000.

Note: Use with caveat: students excluded after methodology change; cross-time comparability to pre-2023 data is limited.

Note: Reported in Oct 2025; cessation occurred in March 2025. Code event_date as public-report date and update_date as approximate cessation month.

Note:

Note: Do not claim 20% of all Sber employees without filing-level confirmation; strongest fact is the public first-person statement and the responsibility dispute.

Note: Different states use different definitions (bell-to-bell vs class-time-only vs district-discretion). Trend direction unambiguous; exact count varies by source taxonomy.

Note: Adds enforcement-quality layer to the existing v0.2 LL144 signal.

Note: Use official FNS-backed counts where possible; media summaries are secondary.

Note: The report’s job-cut attribution is political/compiled evidence; code as policy signal, not official labor statistics.

Note: Use legal text as primary; practitioner summaries for implementation details.

Note: Factcheck adjusted: uploaded candidate dated the signing as 2024; checked sources point to 2025 tax reform / 2026 effective date. Keep confidence medium until official law text is pinned.

Note: Allegations unresolved; code as opacity/privacy/process signal, not proven bias.

Note: Measurement-update signal; does not itself provide realized-transition panel data.

Note: Exact resumes-per-vacancy values are secondary/aggregated; use as directionally robust junior-squeeze signal, not official hh.ru microdata.

Note: Use as frame only: not official labor statistics, not January-only, and AI/automation attribution is based on tracker/reporting categories.