Trump Administration Releases National Policy Framework on Artificial Intelligence

Summary

• On March 20, 2026, pursuant to President Trump’s Executive Order issued on December 11, 2025, titled “Ensuring a National Policy Framework for Artificial Intelligence” (the “National Policy EO”),[1] the White House released a legislative blueprint for a national artificial intelligence (“AI”) policy framework (the “Framework”).[2]
• The Framework urges Congress to adopt a federally unified, innovation-oriented regime centered on preemption of state AI laws and a “light-touch” regulatory approach.
• The Framework recommends targeted federal standards in areas such as child safety, digital replicas, and infrastructure development, while leaving questions as to whether training AI models on copyrighted content violates existing copyright laws or constitutes “fair use,” to ongoing judicial resolution.

The Framework

White House officials state that the Framework “delivers on the [National Policy EO] while expanding workforce and education opportunities so American workers benefit from AI-driven growth.”[3] It sets out legislative recommendations across seven areas:

1. Child safety: Require AI services and platforms to implement safeguards that reduce risks of sexual exploitation and self-harm; provide parents with tools to manage children’s privacy, screen time, and content exposure; establish age-assurance requirements; clarify that existing child privacy laws apply to AI systems; avoid vague content standards that could increase litigation; and preserve state authority to enforce generally applicable child-protection laws.
2. Communities: Promote AI-driven economic growth and infrastructure development while protecting communities by preventing electricity cost increases tied to AI data centers; streamlining federal permitting for AI infrastructure; strengthening enforcement against AI-enabled fraud; enhancing national security capabilities to assess frontier AI risks; and supporting small businesses through grants and tax incentives.
3. Creators: While noting that the Administration believes that training AI models on copyrighted material does not violate copyright laws, it supports deferring to courts to resolve this issue; not taking any action that would impact the judiciary’s resolution of whether training on copyrighted material constitutes fair use; exploring voluntary licensing or collective rights frameworks; establishing safeguards against unauthorized digital replicas of individuals’ voice, likeness, or other attributes; and monitoring legal developments to address protection gaps.
4. Censorship: Safeguard First Amendment protections by prohibiting government coercion of platforms to moderate content based on partisan or ideological viewpoints and by providing effective mechanisms for individuals to seek redress for federal actions that censor or influence lawful expression on AI systems.
5. Competitiveness: Advance U.S. leadership in AI by establishing regulatory sandboxes; expanding access to AI-ready federal datasets for industry and academia; and relying on sector-specific regulators and industry-led standards rather than creating a new overarching AI regulator.
6. Workforce and Education: Ensure workers benefit from AI-driven growth by integrating AI into education and workforce training; expanding research on AI-driven labor market impacts; and strengthening land-grant universities’ capacity for skills development, technical assistance, and youth engagement.
7. Preemption of State AI Laws: Establish a national AI policy that promotes innovation and competitiveness by preempting burdensome state AI laws, while preserving core state authorities (e.g., general law enforcement, zoning, and state AI use) and preventing state restrictions that conflict with national AI strategy.

Overall, the Framework reflects a “light touch” federal regulatory approach that emphasizes innovation, preemption of state AI laws, and reliance on existing legal regimes, including deferring to courts on unsettled questions such as whether training AI models on copyrighted data constitutes fair use. Although legislative adoption remains uncertain, the proposal is likely to shape both near-term congressional activity and longer-term regulatory architecture. In the interim, companies should consider operations in a hybrid regulatory environment while positioning for potential federal standardization.