United States of AmericaAI regulation

Regulation Status

• ❗️ This profile card covers US federal developments only. For state-level developments, find the US state profile card below (or in the tab menu above).
• The US has no single comprehensive, binding federal AI statute; it relies on sectoral and targeted laws (notably the TAKE IT DOWN Act), executive orders and voluntary frameworks. Regulation is fragmented across federal, state and local levels — California and New York lead on volume of proposed bills — and the Trump administration has moved toward a deregulatory, pro-innovation, preemption-oriented stance, with Executive Order 14365 ("Ensuring a National Policy Framework for Artificial Intelligence", December 2025) directing an AI Litigation Task Force within the Department of Justice to challenge state AI laws as unduly burdening interstate commerce.
• Federal AI activity spans national security, bias and fairness, transparency and disclosure, and accountability, pursued through a dense mix of executive orders, agency frameworks and proposed bills rather than a single comprehensive statute; instruments range from strategic and administrative initiatives to rules on specific issues such as deepfakes and foundation models. In June 2026, federal activity also included frontier-model access controls for cybersecurity and national-security reasons: after an export-control directive affecting Anthropic's Fable 5 and Mythos 5, the government later lifted immediate controls on 30 June 2026, with Claude Fable 5 set for global redeployment and the less-safeguarded Mythos 5 restored for select domestic organizations in a defensive cybersecurity program. On 29 September 2026, an executive order directed agencies, to the maximum extent permitted by law, to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in non-statutory executive-branch materials, and directed the preparation of proposed legislative language on a federal definition. The order also directed agencies to integrate public-facing services and websites with the AI-powered America.gov portal, which is to provide responses from official federal sources and meet accuracy, reliability and transparency requirements.
• The TAKE IT DOWN Act was signed into law on 19 May 2025, criminalising the publication of nonconsensual intimate imagery (including AI-generated/deepfake content) and requiring covered platforms to remove such material within 48 hours; it is now being actively enforced, with the FTC issuing compliance guidance and launching a complaint platform and the DOJ bringing its first criminal charges in May 2026.
• On 24 September 2026, senators introduced the Artificial Intelligence Risk Management and Security Act of 2026, which would establish a permanent AI Safety Board in the Department of Commerce and require frontier-model developers to provide access for safety evaluation at least 45 days before public release, as well as file safety plans and report specified incidents. A House bill introduced the same day, the Ban Artificial Superintelligence Act of 2026 (HB 10538), would prohibit development and deployment of artificial superintelligence and specified precursor systems, impose a pause on certain advanced AI development, and establish a cabinet-level Department of Artificial Intelligence. Both are proposals, not enacted laws.