AI law and policy in United States
10 instruments relevant to AI in United States are indexed here, 6 of them binding law, 2 of which name AI or deepfakes in their titles.
Governance readiness 6 of 10 points on the governance readiness index, which records how much machinery exists, not how strict it is. AI Governance Readiness Index This site Our own editorial coding from primary legal sources, not an objective measurement. Every score cites a specific instrument and article. "Readiness" means the machinery exists to govern AI. It is emphatically not a judgement that the resulting rules are good, proportionate, or well enforced. Cite AI Safety Tracker, "AI Governance Readiness Index". https://aisafetytracker.org/methodology/governance-readiness/. Retrieved 23 October 2026.
Use of AI, AI datacentres and governance readiness beside this index: AI in the United States
Statutes
Act of Parliament. Binding on everyone it applies to.
-
Colorado General Assembly
Repealed and re-enacted the 2024 Colorado AI Act, SB24-205, in May 2026, before its duties applied. The impact assessments and risk-management programme are gone; in their place are disclosure duties and rights of access, correction and human review for automated decisions in consequential matters. Developer duties apply from January 2027.
- consumer
- transparency
-
California State Legislature
State law, and the first binding frontier-model transparency statute anywhere: developers above a compute threshold must publish a safety framework and report critical safety incidents. It reaches most frontier labs because most of them are in California.
- safety
- transparency
-
New York State Legislature
Signed in December 2025 and narrowed by a chapter amendment in March 2026 to track California's SB 53: developers with over $500 million in annual revenue that train models above 1e26 FLOP must publish safety protocols and report safety incidents to the state. Applies from January 2027.
- safety
- transparency
-
TAKE IT DOWN Act 2025
US Congress
Criminalises publication of non-consensual intimate imagery including AI-generated depictions, and requires covered platforms to remove it on notice. One of very few federal statutes that names AI-generated material at all.
- safety
- children
-
US Congress
The general prohibition on unfair or deceptive acts and practices. With no AI statute in place it is the main federal hook for AI claims that do not hold up, and the FTC has used it against several AI products.
- consumer
- competition
Rules made under an Act
Rules, regulations or codes made under an Act. Binding.
-
Bureau of Industry and Security, Department of Commerce
The most materially binding AI rules the United States has, and the least discussed as AI policy. They restrict which accelerators can be sold where, which decides who can train large models at all.
- export control
- security
Binding on government only
Binds government agencies only, not the wider market.
-
President of the United States
Revoked Executive Order 14110 and reset federal policy toward removing regulatory barriers. Executive orders bind the executive branch, not the market, and a successor can revoke them in a day, as this one did.
- government use
-
President of the United States
Directs federal agencies to review state AI laws, a Justice Department task force to challenge them in court, and some federal funding to be conditioned on them, pending a federal standard that would preempt state law. It changes no state law by itself, and Congress has passed no preemption statute.
- government use
Voluntary standards
Guidance with no legal force. Compliance is a choice.
-
National Institute of Standards and Technology
No legal force, and the most widely adopted AI governance document in the United States. State bills and procurement rules incorporate it by reference, which is how a voluntary framework acquires teeth.
- safety
- government use
Reviews and inquiries
An inquiry or review. Evidence and recommendations, not obligations.
-
President of the United States
Listed because it is still widely cited as current US policy and is not. It introduced the 1e26 FLOP reporting threshold, the second compute threshold written into any official instrument, and was revoked in January 2025.
- safety
- security
About this index
Compiled here from primary instruments and licensed MIT.
6 of the 10 instruments here are binding law, and 2 of those name AI. The remainder apply to AI without naming it, so a keyword search of the statute book does not find them.
Topics covered: children, competition, consumer, export control, government use, safety, security, transparency. A build check probes every citation on this page, because a link that returns a 404 looks like evidence and is not.
This is not legal advice. It is not a compliance checklist and it is not exhaustive. Commencement dates, proposals and enforcement positions move; the review date is on every record. If something here is wrong, say so and it gets fixed and listed rather than quietly patched.