Rules · the AI law calendar

EU AI Act to require machine-readable marking on older generative AI

2 Dec 2026 is 20 days away, with 1 other duty on the same day. 22 dates in force lie ahead, in 6 laws. As of 12 Nov 2026, from each law’s official text. This page is a factual record of the dates in each law’s official text, quoted from a kept copy. It is not legal advice and does not say whether a law applies to you. Read the law itself, and ask a lawyer.

2 Dec 2026 EU AI Act to require machine-readable marking on older generative AI Providers of systems that make synthetic audio, image, video or text, placed on the market before 2 August 2026, must mark their output as Article 50(2) requires. For: providers of generative AI systems on the market before 2 August 2026.
2 Dec 2026 EU AI Act to ban AI made for non-consensual sexual imagery Two bans that the Digital Omnibus on AI added to Article 5 apply. They cover AI systems that make realistic sexual images, video or audio of an identifiable person without consent, and AI systems that make child sexual abuse material. A system is banned when that is its purpose, or when it can readily do so and lacks adequate safeguards. A deployer must not use a system for that purpose. For: providers and deployers.
1 Jan 2027 California’s AI Transparency Act to bar hosting of AI systems without disclosures A hosting platform must not knowingly offer a generative AI system that does not place the disclosures the Act requires. For: GenAI system hosting platforms.
1 Jan 2027 California’s AI Transparency Act to require provenance checks on large platforms A large online platform must detect standard provenance data in content it distributes. It must show users what provenance data is available and let them inspect it. Where technically feasible, it must not knowingly remove that data. For: large online platforms.
1 Jan 2027 California’s frontier AI law to start yearly whistleblower reports The Attorney General starts a yearly report of anonymized, aggregated information about whistleblower reports from covered employees. For: California Attorney General.
1 Jan 2027 California’s frontier AI law to start yearly reviews of its definitions The Department of Technology must review the Act's definitions of frontier model, frontier developer and large frontier developer each year, and recommend updates. For: California Department of Technology.
1 Jan 2027 California’s frontier AI law to start yearly safety-incident reports The Office of Emergency Services starts a yearly report of anonymized, aggregated information about critical safety incidents. For: California Office of Emergency Services.
1 Jan 2027 Colorado’s AI law to take effect The reenacted law starts. A developer of a covered ADMT must give each deployer documentation of its intended uses, categories of training data, known limitations, and instructions for appropriate use and human review. For: developers of automated decision-making technology.