Regulation
Dated obligations by jurisdiction: what applies, to whom, and from when.
| Jurisdiction | Instrument | Obligation | Applies to | From | Status | Source |
|---|---|---|---|---|---|---|
| China | Interim Measures for the Management of Generative AI Services (2023) | Public-facing generative AI services must use lawful training data, keep output lawful, label AI-generated content, protect users, and (where there is public-opinion capacity) complete a security assessment and algorithm filing. | Providers of generative AI services to the public in mainland China | 2023-08-15 | In force | Cyberspace Administration of China: official textchecked 2026-10-11 |
| European Union | Regulation (EU) 2024/1689 (AI Act), Art. 5 and 4 | Bans unacceptable-risk AI practices (social scoring, untargeted facial-image scraping, manipulative or exploitative systems, most real-time public biometric identification by police) and requires operator AI literacy. | All operators of AI systems | 2025-02-02 | In force | European Commission: AI regulatory framework timelinechecked 2026-10-11 |
| European Union | Regulation (EU) 2024/1689 (AI Act), Ch. V (Art. 53-55) | Transparency, technical-documentation and copyright-policy duties for general-purpose AI models, with extra obligations for the most capable (systemic-risk) models. | Providers of general-purpose AI (GPAI) models | 2025-08-02 | In force | European Commission: AI regulatory framework timelinechecked 2026-10-11 |
| China | Measures for Labeling AI-Generated Synthetic Content (2025) | Providers must add visible labels to AI-generated text, images, audio, video and virtual scenes and embed metadata labels in the files; distribution platforms must detect and flag AI-generated content. | Providers of AI content-generation services and distribution platforms | 2025-09-01 | In force | Cyberspace Administration of China: official textchecked 2026-10-11 |
| United States, Texas | HB 149 (2025), Responsible AI Governance Act (TRAIGA) | Bans AI built to incite self-harm or crime, to unlawfully discriminate against a protected class, for certain government social scoring, or for unconsented government biometric identification; government must disclose AI interactions. | Developers and deployers of AI (some duties specific to government agencies) | 2026-01-01 | In force | Texas Legislature: HB 149 enrolled textchecked 2026-10-11 |
| United States, California | AB 2013 (2024), Generative AI Training Data Transparency | Developers must post a high-level summary of the datasets used to train a generative AI system, including sources, time periods, and whether it includes personal, copyrighted or synthetic data. | Developers of generative AI made available to Californians | 2026-01-01 | In force | California Legislature: AB-2013 textchecked 2026-10-11 |
| United States, California | SB 53 (2025), Transparency in Frontier AI Act | Large frontier developers must publish an AI safety framework and transparency reports, report critical safety incidents, and send catastrophic-risk summaries to state emergency services; whistleblower protections apply. | Large frontier AI developers (models above 10^26 FLOP, over $500M affiliated revenue) | 2026-01-01 | In force | California Legislature: SB-53 textchecked 2026-10-11 |
| South Korea | AI Basic Act (Act No. 20676) | National AI governance regime: transparency and labeling for generative AI output, risk-management and safety duties for high-impact AI, and a domestic-representative requirement for large overseas providers. | AI businesses operating in Korea, including certain foreign providers | 2026-01-22 | In force | Korea Law Information Center: Framework Act on AIchecked 2026-10-11 |
| United States (federal) | TAKE IT DOWN Act, Public Law 119-12 (2025) | Covered platforms must run a notice-and-removal process and take down reported non-consensual intimate images, including AI deepfakes, within 48 hours of a valid request. | Online platforms that host user-generated content | 2026-05-19 | In force | Public Law 119-12 (GovInfo)checked 2026-10-11 |
| European Union | Regulation (EU) 2024/1689 (AI Act), Art. 50 | Transparency duties: tell people when they are interacting with an AI, and label AI-generated or manipulated content such as deepfakes. General application date of the Regulation. | Providers and deployers of AI that interacts with people or generates synthetic content | 2026-08-02 | In force | European Commission: AI regulatory framework timelinechecked 2026-10-11 |
| United States, Colorado | SB 26-189 (2026), Automated Decision-Making Technology | Developers of automated decision-making tech used in consequential decisions must give deployers documentation for required disclosures; deployers must give pre-use notice and post-adverse-outcome explanations. | Developers and deployers in consequential decisions (employment, housing, finance, insurance, healthcare, education, benefits) | 2027-01-01 | Upcoming | Colorado General Assembly: SB 26-189checked 2026-10-11 |
| European Union | Regulation (EU) 2024/1689 (AI Act), Annex III, as amended by the 2026 Digital Omnibus | Full duties for high-risk AI (biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration): risk management, data governance, logging, human oversight, conformity assessment and registration. | Providers and deployers of standalone high-risk AI systems | 2027-12-02 | Upcoming | European Commission: AI regulatory framework timelinechecked 2026-10-11 |