Wiki · Enforcement timeline
Catalog dashboard
Enforcement timeline
Chronologically-ordered enforcement actions linked to the instruments and topics tracked in the Policy Window catalog. Closed cases sort by their resolution year; ongoing cases sort by initiation. Each entry links to the catalog topics + the catalog instruments the case touches, and (where possible) to the best-available primary source for the action.
13 catalogued cases total · 3 match the current filter (year = 2024; topic = Individual Redress) clear all.
Filters
Cases (3)
- Settlement with remedy
- Withdrawn
- Open arrow = ongoing (no resolution year yet)
- Settlement with remedy
2023 → 2024 · EU
EDPB ChatGPT Taskforce
Enforcer: European Data Protection Board (EDPB) — coordinated DPA action
Target: OpenAI
Violation alleged: Italian Garante temporarily banned ChatGPT (Mar-Apr 2023) over alleged lack of legal basis for training-data processing, missing age-verification, and inability to honour data-subject rights. EDPB convened taskforce to coordinate DPA responses.
Lesson: First EU-wide AI enforcement coordination predating the EU AIA. Established that GDPR applies fully to LLM training + deployment + that DPAs would coordinate via EDPB rather than fragment. ChatGPT resumed Italian service after age-verification + Article-15 right-of-access endpoint additions. Direct precedent for EU AIA Art. 53 implementation timeline.
- Withdrawn
2023 → 2024 · IN
MEITY deepfake takedown advisories
Enforcer: Ministry of Electronics and Information Technology (MEITY)
Target: Multiple intermediaries — Meta, YouTube/Google, X, several Indian social platforms
Violation alleged: Failure to take down political deepfake content within statutory windows (36 hours under IT Rules 2021). MEITY's Mar-2024 advisory additionally required pre-deployment-approval for AI models above unspecified capability thresholds; rescinded Apr-2024 after frontier-lab pushback.
Lesson: India's compressed legislative cycle: a sweeping pre-deployment-approval requirement (closer to CN registration than US sectoral) was rescinded within 5 weeks after industry + civil-society pushback. Demonstrates that Global South AI regulation is in active design AND that even nationally-coercive states face frontier-lab leverage. Indian regulatory approach now favours post-deployment incident reporting + IT-Rules takedown.
Source: https://www.meity.gov.in/writereaddata/files/Advisory%201%20March%202024.pdf
- Settlement with remedy
2024 · US
Texas AG v. Pieces Technologies (healthcare AI deceptive practices)
Enforcer: Texas Attorney General
Target: Pieces Technologies, Inc.
Violation alleged: Pieces marketed its generative-AI clinical-summary product to Texas hospitals with materially misleading accuracy claims (alleging severe-hallucination rates orders of magnitude lower than measured in deployment).
Lesson: First US state-attorney-general action against a generative-AI vendor under state UDAP/UDTPA authority. Settlement (Sep 2024) required: (1) clear disclosure of AI involvement to end-users, (2) accurate marketing-claim substantiation, (3) ongoing monitoring of model output for hallucination rates. Establishes that state AGs can reach AI deployment claims without an AI-specific state statute — through long-standing consumer-protection authority. Likely template for parallel actions by other state AGs.
Source: https://www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-secures-settlement-healthcare-technology-company-pieces-technologiesregulator landing
Editorial scope
This timeline reflects enforcement cases tracked in the Policy Window catalog. It is not exhaustive; coverage focuses on high-precedent matters relevant to the catalogued instruments (currently 13 cases across EU, US, UK, China, India, Italy, France). The catalog deliberately omits routine regulator letters and ongoing investigations whose materials are not public.
For the inclusion rubric (when a case enters the catalog, what level of source-defensibility is required, how jurisdictional balance is managed), see /wiki/methodology. Cases marked “ongoing” remain editorial-watch items; outcomes get backfilled as the public record settles.