United States
United States AI policy combines federal executive action, agency standards, congressional proposals and state law.
Policy approach
Current structureNo single horizontal federal AI statute governs the entire field. Congress, federal agencies and states advance overlapping rules on disclosure, evaluation, infrastructure and sectoral use.
Policy questions
Should AI-generated content be labelled?
Governments require disclosure in different forms. Technical duties, responsible actors and enforcement models still diverge.
Should frontier models face mandatory evaluations?
Evaluation language is widespread in policy texts. Thresholds, evaluators and legal consequences remain contested.
Should AI data centres face growth controls?
Electricity, water, local consent and national competitiveness are turning compute infrastructure into a policy fight.
Should intergovernmental AI principles guide national rules?
Soft-law instruments such as the OECD AI Recommendation supply shared definitions that many jurisdictions then incorporate, revise or ignore.
Should states implement UNESCO’s AI ethics recommendation?
UNESCO’s 2021 recommendation is a global ethics standard; the open policy question is how far states turn its principles into domestic duties.
Should states join binding AI human-rights treaties?
The Council of Europe Framework Convention is open for signature. The live dispute is ratification and domestic effect, not whether a treaty text exists.
Should advanced AI chips face export controls?
US export controls already treat advanced computing integrated circuits as regulated items. The dispute is scope, partners and effectiveness.
Should states prefer AI promotion statutes to horizontal bans?
Japan’s Act No. 53 of 2025 promotes AI development and directs guidelines. Whether that model outcompetes bans-heavy statutes remains contested.
Should watermarking standards be mandatory?
Open provenance standards such as C2PA already exist. The dispute is whether law should require them or leave adoption voluntary.
Should sector regulators own AI rules instead of a horizontal act?
The UK white paper and Singapore’s Model Framework both privilege existing regulators or voluntary tools over an early omnibus AI Act.
Policy asks
Require public disclosure rules for foundation-model training data, testing and operations
Direct a federal regulator to establish public-information requirements covering training data, documentation, testing, inference-time collection and model operations for covered foundation models.
Use voluntary AI risk-management frameworks for organisational governance
Adopt voluntary frameworks such as the NIST AI RMF to manage trustworthiness risks without treating the framework text as a statute.
Pause new large AI data centres pending resource and security review
Pause construction or upgrading of covered AI data centres until Congress enacts specified safeguards and expressly terminates the moratorium.
Publish authoritative AI data-centre electricity demand projections
Maintain public, method-stated projections of data-centre and AI-focused electricity demand for grid and infrastructure planning.
Adopt the OECD AI-system definition and trustworthy-AI principles
Use the OECD Recommendation’s AI-system definition and principles as a shared baseline for domestic policy and legislative drafting.
Control advanced computing integrated circuits under export administration rules
Use Export Administration Regulations revisions to licence or restrict advanced computing ICs used in AI datacentre deployments.
Adopt open content-provenance standards such as C2PA Content Credentials
Use open technical specifications for content provenance and authenticity assertions in synthetic-media governance.
Policy vehicles
H.R. 8094 · AI Foundation Model Transparency Act of 2026
A House bill directing the FTC to establish public disclosure requirements for covered foundation models.
H.R. 9442 · Artificial Intelligence Data Center Moratorium Act
A House bill proposing a conditional moratorium on construction or upgrading of covered AI data centres.
NIST AI Risk Management Framework
US federal voluntary framework for managing AI trustworthiness risks across design, development, use and evaluation.
BIS advanced computing IC interim final rule (Jan 2025)
US EAR revision controlling advanced computing integrated circuits relevant to AI datacentre chips.
C2PA Content Credentials technical specification
Open technical standard for attaching and validating content provenance assertions.
Policy timeline
Events and deadlines- 2023-01-26NIST releases AI Risk Management Framework 1.0
NIST published the voluntary AI Risk Management Framework.
- 2025-01-01C2PA publishes technical specification 2.3
C2PA Specification 2.3 defines the Content Credentials provenance standard.
- 2025-01-16BIS advanced computing interim final rule takes effect
The Federal Register interim final rule revising EAR advanced computing IC controls took effect.
- 2025-09-29California SB 53 chaptered into law
SB 53 was approved by the Governor and chaptered as Chapter 138, Statutes of 2025.
- 2026-01-01California frontier AI transparency duties take effect
Chapter 25.1 of the Business and Professions Code became effective.
- 2026-03-26AI Foundation Model Transparency Act introduced in the House
H.R. 8094 entered the 119th Congress and was referred for committee consideration.
- 2026-06-24AI Data Center Moratorium Act introduced in the House
H.R. 9442 was introduced and referred to House committees.