Countries / United States (federal)

United States (federal)

No comprehensive federal AI statute: Washington governs AI through executive orders, OMB rules for federal agencies, export controls and sector regulators, while the White House and DOJ try to preempt or litigate against state AI laws.

PatchworkApproach
15Rules tracked
9Binding and in force
1Upcoming deadlines

In force (12)

  • EO 14179
    Executive orderThe President (Executive Office of the President)Binding

    Revokes the Biden-era EO 14110 on safe, secure and trustworthy AI and directs agencies to review and rescind actions taken under it. Sets federal policy to 'sustain and enhance America's global AI dominance' and ordered the AI Action Plan delivered in July 2025.

    In forceEffective 23 Jan 2025
  • AI Action Plan
    StrategyExecutive Office of the President (OSTP, Special Advisor for AI and Crypto, NSC)Non-binding

    Federal AI strategy built on three pillars (accelerate innovation, build AI infrastructure, lead in international AI diplomacy and security) that directs agencies to remove regulatory barriers, revise the NIST AI RMF, expand CAISI model evaluations, speed data-center permitting and promote exports of the US 'AI stack'. It also tells agencies to weigh a state's AI regulatory climate when awarding AI-related funding.

    In forceEffective 23 Jul 2025
  • State AI law preemption EO
    Executive orderThe President (Executive Office of the President)Binding

    Orders the Attorney General to set up an AI Litigation Task Force to challenge state AI laws, tells Commerce to publish a list of 'onerous' state AI laws within 90 days and to bar those states from remaining BEAD broadband non-deployment funds, and asks the FTC (policy statement) and FCC (possible preemptive disclosure standard) to act. It also orders a legislative proposal for a preemptive federal framework that would spare state child-safety, data-center and state-procurement laws.

    In forceEffective 11 Dec 2025
  • EO 14409 (covered frontier models)
    Executive orderThe President (Executive Office of the President)Binding

    Directs NSA, CISA, Treasury and NIST to build a classified benchmark for the cyber capabilities of AI models and a threshold for designating 'covered frontier models', plus a voluntary framework under which developers give the government up to 30 days' pre-release access to such models. It also orders CISA directives on AI-enabled cyber defence, a Treasury-led AI vulnerability clearinghouse, and DOJ prioritisation of prosecutions for AI-enabled hacking.

    In forceEffective 2 Jun 2026
  • EO 14434 ('Super Intelligence' terminology)
    Executive orderThe President (Executive Office of the President)Binding

    Requires executive-branch agencies, to the extent permitted by law, to use 'Super Intelligence' and 'SI' instead of 'Artificial Intelligence' and 'AI' in correspondence, websites, reports and policy documents, defining SI by reference to the existing statutory definition of AI (15 U.S.C. 9401(3)). It asks OSTP to propose legislation within 60 days on whether a statutory 'Super Intelligence' definition should supersede 'artificial intelligence'.

    In forceEffective 29 Sep 2026
  • OMB M-25-21
    GuidelineOffice of Management and Budget (OMB)Binding

    Binding guidance for all federal agencies (including independent regulators) that replaces M-24-10: agencies must appoint Chief AI Officers, publish AI strategies and use-case inventories, and apply minimum risk-management practices (pre-deployment testing, impact assessments, human oversight) to 'high-impact AI'. Agencies had 365 days (to 3 Apr 2026) to document those practices and must stop using non-compliant high-impact AI.

    In forceEffective 3 Apr 2025
  • OMB M-25-22
    GuidelineOffice of Management and Budget (OMB)Binding

    Replaces M-24-18 and sets rules for how agencies buy AI: preference for American AI, protections against vendor lock-in, limits on vendors' use of government data, and performance-based contracting. It applies to contracts awarded under solicitations issued 180 days or more after issuance and to options exercised after that date.

    In forceEffective 30 Sep 2025
  • OMB M-26-04 ('Woke AI' procurement rules)
    GuidelineOffice of Management and Budget (OMB)Binding

    Implements EO 14319 (Preventing Woke AI in the Federal Government) by requiring agencies to buy only large language models that meet two 'Unbiased AI Principles' — truth-seeking and ideological neutrality — and to write compliance terms and vendor disclosure requirements into LLM contracts. Agencies had to update procurement policies by 11 Mar 2026 and should amend existing LLM contracts before exercising options.

    In forceEffective 11 Dec 2025
  • BIS H200 case-by-case rule
    RegulationBureau of Industry and Security (BIS), Department of CommerceBinding

    Changes BIS licence review for exports to China and Macau of Nvidia H200-class and less advanced AI chips from a presumption of denial to case-by-case review, if the exporter certifies sufficient US supply, no diversion of foundry capacity, adequate recipient security, and independent US third-party performance testing. More advanced chips remain under the existing controls.

    In forceEffective 15 Jan 2026
  • TAKE IT DOWN Act
    LawU.S. CongressBinding

    Makes it a federal crime to knowingly publish non-consensual intimate images, including AI-generated 'digital forgeries', and threats to do so. Since 19 May 2026 'covered platforms' must run a notice-and-removal process and take down reported images and known identical copies within 48 hours, enforced by the FTC.

    In forceEffective 19 May 2025
  • FDA PCCP guidance
    GuidelineU.S. Food and Drug Administration (CDRH, CBER, CDER)Non-binding

    Final FDA guidance explaining how makers of AI-enabled medical devices can include a Predetermined Change Control Plan in 510(k), De Novo or PMA submissions describing planned model modifications, the protocol for developing and validating them, and an impact assessment. Changes made within an authorised PCCP do not need a new marketing submission.

    In forceEffective 4 Dec 2024
  • NIST AI RMF
    StandardNational Institute of Standards and Technology (NIST)Non-binding

    Voluntary framework (Govern, Map, Measure, Manage) for managing AI risks across the lifecycle, with a Generative AI Profile and, since 7 Apr 2026, a concept note for a critical-infrastructure profile. NIST states the AI RMF 1.0 'is being revised as part of the White House AI Action Plan', which calls for removing references to misinformation, DEI and climate change.

    In forceEffective 26 Jan 2023

Proposed or in consultation (3)

  • White House AI legislative framework
    StrategyThe White HouseNon-binding

    Non-binding recommendations asking Congress to pass a uniform federal AI framework that preempts 'unduly burdensome' state AI laws while preserving state police powers, generally applicable child-protection laws and zoning. It also proposes age-assurance and anti-self-harm features for AI services used by minors, a federal digital-replica right, regulatory sandboxes, and no new federal AI regulator.

    ProposedProposed 20 Mar 2026
  • Great American AI Act
    BillU.S. House of Representatives — Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA)Non-bindingPartly confirmed

    Bipartisan discussion draft that would require large frontier AI developers to publish safety frameworks and transparency reports, report critical safety incidents to NIST's Center for AI Standards and Innovation, and undergo periodic independent audits. It would preempt state laws that specifically regulate AI model development for three years, while leaving state laws on AI use and deployment in place.

    ProposedProposed 4 Jun 2026
  • FTC AI accuracy policy statement
    GuidelineFederal Trade Commission (FTC)Non-binding

    Proposed FTC policy statement on how Section 5's ban on deceptive practices applies to companies marketing AI systems, focusing on undisclosed steering or distortion of AI outputs. It takes the position that complying with a state law does not excuse undisclosed distortion of outputs and that state laws requiring such conduct may be preempted.

    In consultationProposed 7 Jul 2026

What the rules require

Obligation types found across United States (federal)'s instruments. Filled dot: imposed by a binding instrument. Ring: guidance only.

Risk classificationOMB M-25-21, NIST AI RMF
Prohibited usesTAKE IT DOWN Act
Transparency to usersGreat American AI Act, OMB M-26-04 ('Woke AI' procurement rules), FDA PCCP guidance, FTC AI accuracy policy statement
Registration or filingBIS H200 case-by-case rule
Impact assessmentOMB M-25-21, FDA PCCP guidance, NIST AI RMF
Human oversightOMB M-25-21, NIST AI RMF
Incident reportingGreat American AI Act
Data governanceOMB M-25-22, FDA PCCP guidance, NIST AI RMF
Frontier model safetyGreat American AI Act, EO 14409 (covered frontier models)
Security testingAI Action Plan, Great American AI Act, EO 14409 (covered frontier models), BIS H200 case-by-case rule
Government use of AIAI Action Plan, EO 14434 ('Super Intelligence' terminology), OMB M-25-21, OMB M-25-22, OMB M-26-04 ('Woke AI' procurement rules)
Consumer rightsTAKE IT DOWN Act
PenaltiesGreat American AI Act, TAKE IT DOWN Act

Timeline

States we track

California, Colorado, Texas, New York, Illinois, Utah