Trump drafts order to preempt state AI laws, centralize rules

A draft in circulation directs Attorney General Pam Bondi to create a Justice Department task force to challenge state measures deemed inconsistent with federal policy, restrict certain federal funds to states with “onerous” rules, and request the Federal Trade Commission to issue guidance that prioritizes federal consumer protection standards over conflicting state regulations.

The document outlines a plan to centralize AI governance in Washington and reduce state-by-state requirements. AI and crypto czar David Sacks would oversee implementation. The text remains under review and could change before any release.

Under the draft, the Commerce Secretary would assess states’ eligibility for specified broadband funding within 90 days based on their AI statutes. It would direct FCC Chair Brendan Carr to begin proceedings to set federal AI disclosure standards that would supersede conflicting state rules. Agencies would also examine whether funding programs should be conditioned on states adopting policies aligned with federal standards.

California is highlighted. The draft criticizes SB 53 for placing complex risk disclosure mandates on major AI developers. It also references SB 243, a separate law governing AI companion chatbots that requires them to self‑identify and restrict certain content for minors.

On Capitol Hill, House Republicans are weighing language to preempt state AI laws in the National Defense Authorization Act before year‑end, with Majority Leader Steve Scalise telling Punchbowl News they are “looking at” inserting the provision into the defense bill. In July, the Senate voted 99-1 to remove a 10-year moratorium on state AI laws from a broader package after Senator Marsha Blackburn withdrew support for a five-year compromise reached with Senate Commerce Chair Ted Cruz.

Trump has argued that fragmented state regimes could slow AI progress. In a post on Wednesday, Nov. 19, 2025, he wrote that some states are “trying to embed DEI ideology into AI models, producing ‘Woke AI,’” and added, “We MUST have one Federal Standard instead of a patchwork of 50 State Regulatory Regimes. If we don’t, then China will easily catch us in the AI race.”

Legal and industry voices are watching how preemption might work in practice. Sharon Klein, a partner at Blank Rome and co-chair of the firm’s Privacy, Security & Data Protection practice, viewed a federal baseline as potentially helpful for compliance in states without AI statutes. She cautioned that a single rulebook could slow rapid responses to AI-related harms at the local level, with the impact hinging on how any mandate is implemented, funded, and enforced.

The draft envisions a DOJ task force to identify and challenge state AI measures that conflict with federal policy. The FTC would be asked to clarify how existing consumer protection laws can police AI practices nationwide and, where applicable, preempt overlapping or inconsistent state rules. The text favors national disclosure and safety frameworks over state-by-state obligations for risk reporting, safety evaluations, or developer certifications.

The draft cites the introduction of over 1,000 AI bills in state legislatures, warning that a patchwork of rules could undermine innovation. It envisions a DOJ task force to identify and challenge state AI measures that conflict with federal policy, and asks the FTC to clarify how existing consumer‑protection laws can police AI practices nationwide and, where applicable, preempt overlapping or inconsistent state rules. The text favors national disclosure and safety frameworks over state‑by‑state obligations for risk reporting, safety evaluations or developer certifications.

Key enforcement details remain unresolved. The draft does not define what counts as an “onerous” state rule, how conflicts would be adjudicated, or the timeline for agencies to finalize guidance and standards. Further revisions are expected during the interagency review process.

As we reported earlier, on November 7, 2025, the administration expanded export controls by blocking Nvidia from selling its scaled-back B30A AI accelerator to mainland China, citing clustering risks. Around the same time, Chinese authorities instructed state-funded data centers to rely on domestic AI chips. The actions tightened restrictions that already kept Nvidia’s top Blackwell parts out of China.