Robots Are Better Than People home
Subscribe

What AI policy is coming, and when

AI rules are arriving in pieces: a state here, a regulation there. These are the confirmed dates we are tracking, with links to the official sources.

Key dates

DateWhereWhat starts
Jan 1, 2027ColoradoSB 26-189: duties for automated decision-making technology used in consequential decisions; consumer right to correct inaccurate data
Jan 1, 2027ColoradoHB 26-1263 Chatbot Safety Act: AI disclosure, age estimation, teen protections, self-harm protocols
Dec 2, 2027European UnionAI Act rules for high-risk systems listed in Annex III (as delayed by the AI Omnibus)
Aug 2, 2028European UnionAI Act rules for high-risk AI embedded in regulated products (Annex I)

Already in force

  • EU: the AI Act's bans on certain practices have applied since February 2, 2025, and general-purpose AI model obligations since August 2, 2025. The AI Omnibus, in force from July 27, 2026, also added a ban on AI systems that generate non-consensual sexually explicit content or child sexual abuse material.
  • United States: the TAKE IT DOWN Act (2025) targets non-consensual intimate images, including AI-generated ones.

What to do now

  • Businesses using AI in hiring, lending, housing or insurance decisions: inventory those tools and request vendor documentation.
  • Anyone running a chatbot: make sure it discloses that it is AI and does not claim to be a licensed professional.
  • Selling into Europe: map your systems against the AI Act categories now; the delayed dates still arrive.

We update this page as rules are finalized. Read our full story on Colorado's new law.

Sources and official resources

  1. Colorado Attorney General: ADMT and Chatbot Safety rulemaking
  2. European Commission: AI Omnibus enters into force (Jul 27, 2026)
  3. EU AI Act Service Desk: implementation timeline
  4. Congress.gov: S.146, TAKE IT DOWN Act

Updated November 17, 2026. This guide is general information, not legal or professional advice.

The White-Hat Brief

Weekly: what the machines learned, and how to stay ahead

One email a week. Verified stories, plain-language defenses, no hype. Unsubscribe anytime.