The Trump administration released a seven-point AI policy framework on Friday, March 20, 2026, calling on Congress to preempt state artificial intelligence laws and shield AI developers from what the White House called “open-ended liability.” The blueprint, authored by AI czar David Sacks and Office of Science and Technology Policy director Michael Kratsios, represents the administration’s most detailed attempt yet to shape federal AI legislation.
- The White House released its AI legislative framework on March 20, 2026, calling on Congress to override state AI laws and create a single federal standard.
- Only two states currently have comprehensive AI laws, California’s SB 53 and New York’s RAISE Act, both of which require safety reporting and whistleblower protections.
- An industry coalition including Amazon, Anthropic, Google, Meta, Microsoft, and OpenAI welcomed the framework’s copyright position that AI training on copyrighted material does not violate existing law.
What the White House AI Framework Proposes
The framework lays out seven legislative priorities for Congress: protecting children and empowering parents, safeguarding communities, respecting intellectual property and creator rights, preventing censorship, enabling innovation, developing an AI-ready workforce, and establishing federal preemption of state regulations. Sacks announced the blueprint on social media, calling it a response to “a growing patchwork” of state AI regulations that he said threatened innovation.
On child safety, the framework asks Congress to require “commercially reasonable, privacy protective, age-assurance requirements” for AI platforms likely to be accessed by minors, according to the document. It also calls for measures to combat AI-enabled scams and protect minors from sexual exploitation.
On copyright, the administration took a position that will be closely watched by founders building AI-powered businesses. The framework states that training AI models on copyrighted material does not automatically violate copyright law, and it recommends letting courts resolve the legal disputes between creators and tech companies rather than passing new legislation.
State Law Preemption Is the Most Contested Provision
The most significant provision for startups is the call to block states from regulating AI development directly. Under the framework, states would be prohibited from penalizing AI developers for third-party misuse of their products. States would keep authority over general consumer protection laws, data center zoning, and procurement decisions.
Currently, only two states have comprehensive AI laws on the books. California’s SB 53 and New York’s RAISE Act both mandate whistleblower protections, safety reporting requirements, and transparency about model testing, according to NBC News reporting on the framework. The administration has also actively opposed a proposed Utah bill that would have required AI companies to disclose child protection measures.
The preemption push has failed before. The provision was removed from the GOP budget reconciliation process and was never officially included in the annual defense policy bill, according to Roll Call. More than 50 Republican lawmakers have publicly criticized the White House’s pressure campaign against state AI legislation, stating that “recent attempts to halt state AI legislation suggest not merely a desire for coordination, but an effort to prevent the passage of measures holding the tech industry accountable.”
Industry and Political Reactions Split Along Familiar Lines
A coalition called AI Progress, whose members include Amazon, Anthropic, Google, Meta, Microsoft, Midjourney, and OpenAI, welcomed the framework’s copyright language supporting court resolution of AI training disputes, according to PBS News. Patrick Hedger of the industry group NetChoice said the framework “shows what is at stake and what it will take to win the future.”
Critics pushed back sharply. Rep. Josh Gottheimer called the framework a path to making “the AI industry the Wild West.” Brad Carson of Americans for Responsible Innovation described it as “another chance for tech companies to launch harmful products with no accountability.” House Republican leaders endorsed it immediately, while Senator Marsha Blackburn called it a “roadmap,” according to NBC News.
What the Framework Does Not Address
The blueprint leaves several major questions unanswered. It does not address catastrophic AI risks from out-of-control autonomous systems, large-scale workforce displacement, open-source AI regulation, or national security concerns. For founders building AI businesses, the framework signals a federal preference for light regulation, but its path through Congress remains uncertain given bipartisan resistance to the preemption language.
The framework builds on an executive order signed in December and marks the clearest signal yet from the administration about what kind of AI legislation it would support. Whether Congress moves on any of its seven pillars in the current session will depend on whether the preemption debate can be resolved, a fight that has stalled similar efforts twice before.
Frequently Asked Questions
What Is the White House AI Framework?
The White House AI framework is a seven-point legislative blueprint released on March 20, 2026, by AI czar David Sacks and OSTP director Michael Kratsios. It urges Congress to preempt state AI laws, limit developer liability, and establish federal rules on child safety, copyright, and workforce training.
How Would the AI Framework Affect State AI Laws?
The framework asks Congress to prohibit states from regulating AI development or penalizing AI developers for third-party misuse of their products. States would retain authority over general consumer protection laws, data center zoning, and state procurement decisions.
What Does the AI Framework Say About Copyright and AI Training?
The Trump administration’s framework states that training AI models on copyrighted material does not automatically violate copyright law. It recommends letting courts resolve the issue rather than having Congress pass new legislation restricting AI training data.
