Trump Unveils National AI Framework, Calls for Federal Override of All State AI Laws

The Trump administration unveiled its National Policy Framework for Artificial Intelligence on March 20, 2026, calling on Congress to pass a unified federal AI law 'this year' and outlining seven pillars it wants enshrined in legislation — from child protection to a sweeping ban on states passing their own AI rules. The document, developed by AI czar David Sacks and White House science adviser Michael Kratsios, represents the administration's first comprehensive attempt to shape how artificial intelligence is governed in the United States.

Among the framework's most consequential provisions is a call for sweeping federal preemption of state AI laws. If enacted, it would prevent states like California — which have been leading the charge on AI regulation — from passing their own rules governing the technology. More than 30 states have enacted or are in the process of passing AI legislation; the White House's proposal would override all of them with a single federal standard. Critics, including more than 50 Republican members of Congress, have publicly objected to the preemption push, arguing it amounts to a handout to Silicon Valley. 'This is Washington choosing tech billionaires over American families,' said Rep. Maria Salazar of Florida, one of the Republicans to sign the objection letter. White House officials pushed back, insisting that a fragmented regulatory landscape would cede AI leadership to China.

The framework takes a notably light-touch approach to AI developer liability. It explicitly opposes 'open-ended liability' for AI companies, a position that aligns with heavy lobbying from major developers including Microsoft, Google, and OpenAI. On the politically charged question of copyright and AI training data, the administration stated its belief that AI companies scraping copyrighted content from the internet does not violate U.S. copyright law — a position that has drawn immediate anger from publishers, recording artists, and authors. The Authors Guild called the statement 'an outright surrender to corporate interests at the expense of human creativity.' Several publishers said they planned to challenge the administration's copyright interpretation in court.

Child safety is highlighted as a first-order priority. The framework calls on platforms accessible to minors to implement age-assurance measures and parental account controls, and mandates features designed to prevent sexual exploitation and stop platforms from serving self-harm or eating disorder content to young users. On energy, the framework calls for streamlined permitting for data center power generation while insisting that residential electricity ratepayers should not bear the cost of new AI infrastructure. The free speech pillar directs that AI systems may not suppress lawful political expression and bars government coercion of tech providers on content moderation decisions — a provision critics said would entrench partisan content curation on major platforms.

The broader legislative path remains deeply uncertain. Democrats have signaled they will oppose the preemption language, and a significant bloc of Republicans have their own concerns about protecting state sovereignty. Senate Majority Leader John Thune said the Senate Commerce Committee would hold hearings on the framework in April, though he stopped short of committing to a vote timeline. No bipartisan bill has yet been drafted. Legal experts said the preemption provision is likely to face constitutional challenges, since it would strip states of regulatory authority in an area Congress has never previously occupied. 'Congress can preempt state law,' said Columbia law professor Timothy Wu, 'but the question is whether it can do so in a field it hasn't regulated, by simply declaring that states can't act either.' Technology industry groups broadly welcomed the framework, calling it 'a smart foundation for American competitiveness' and a sign that the administration would not follow Europe's path of comprehensive AI regulation.