March 10, 2026
Federal AI preemption deadline as Commerce names "onerous" state laws
Trump's Dec. 2025 AI executive order set a March 11 deadline for the report
March 10, 2026
Trump's Dec. 2025 AI executive order set a March 11 deadline for the report
On Dec. 11, 2025, President Trump signed an executive order titled Ensuring a National Policy Framework for Artificial Intelligence. It directed the Commerce Department to publish within 90 days a list of existing state AI laws it considers onerous and incompatible with federal policy. The 90-day deadline fell on March 11, 2026. Laws that require AI models to alter truthful outputs or that compel disclosures the administration views as First Amendment violations are specifically targeted in the order's text.
The executive order created three distinct pressure mechanisms working simultaneously. The DOJ's AI Litigation Task Force, established by AG
Pam Bondi on Jan. 9, 2026, can challenge state AI laws in federal court on Commerce Clause or preemption grounds. The Commerce Department can make states with designated laws ineligible for the non-deployment portion of the $42 billion BEAD broadband funding program that Congress appropriated. All federal agencies are directed to review their discretionary grant programs and condition funding on states not enforcing AI laws that conflict with the order.
Legal analysts expected the Commerce report to most likely target Colorado's AI Act, which requires companies deploying high-risk AI systems to take reasonable care to prevent algorithmic discrimination. California's AI transparency statutes, including SB 53 and AB 2013, and New York's RAISE Act, signed Dec. 19, 2025, were also expected targets.
The executive order carves out certain state laws: those protecting children from AI-related harms, state AI procurement rules, and AI infrastructure regulations. Legal scholars noted the carve-out language is ambiguous. The Attorney General retains discretion to designate even child safety or infrastructure laws as onerous if viewed as overly burdensome. The boundaries of the exemption would not be clear until the Commerce report was published and the DOJ task force began filing cases.
The FTC was separately directed to issue a policy statement by March 11, 2026, classifying state-mandated bias mitigation requirements as a per se deceptive trade practice under Section 5 of the FTC Act. The administration's theory is conflict preemption: that state laws requiring companies to adjust AI outputs to reduce bias contradict the FTC's position that such adjustments are themselves deceptive.
If the FTC adopted that position, it would allow the agency to effectively sue states for enforcing their own consumer protection laws. Legal analysts said the theory faces significant obstacles. States have long enforced consumer protection laws affecting how businesses operate, and federal preemption of state consumer protection authority requires a much clearer statutory basis than an FTC policy statement can provide. The statement would still create regulatory uncertainty that chills state enforcement while the legal questions are litigated.
Governors in California, Colorado, and New York stated publicly that the executive order would not stop them from passing or enforcing their AI laws. Rep. Don Beyer, co-chair of the bipartisan Congressional AI Caucus, called the order a terrible idea and said he was exploring legislative responses with members of both parties. Congress had already rejected AI preemption moratorium proposals twice, before H.R. 1 and before the National Defense Authorization Act. The executive order is a workaround of congressional inaction, not a congressional mandate.
Colorado had delayed its AI Act's effective date from Feb. 1 to June 30, 2026. Analysts disagreed about whether the delay resulted from the executive order's pressure or from other legislative factors the Colorado legislature was managing at the same time.
Companies currently operating under state AI laws face a legal reality the executive order cannot change on its own. They remain obligated to comply with those laws regardless of what the Commerce report says. The report designates targets. It does not invalidate statutes. Actual preemption requires either a court injunction blocking a specific state law or a new act of Congress.
The gap between the administration's stated ambitions and what executive power can actually accomplish without Congress is the central tension in the AI preemption story. The administration can pressure, litigate, and cut funding. It cannot unilaterally nullify laws passed by state legislatures. Every state law on the Commerce report's list would have to be challenged individually, in court, under standards of preemption that have historically favored state authority over executive orders.
The Supremacy Clause of the U.S. Constitution establishes that federal law supersedes conflicting state law, but only when Congress has actually enacted the federal law in question. Courts apply two main doctrines: express preemption, where Congress explicitly states it is displacing state law, and implied preemption, which courts infer from the structure and purpose of a federal statute. An executive order is not an act of Congress. It can direct federal agencies but cannot by itself preempt state law.
The administration's strongest legal tool is the spending power, upheld in South Dakota v. Dole (1987), which allows Congress to attach conditions to federal grants. But even that power has limits after NFIB v. Sebelius (2012), which held that conditions become coercive when they threaten to withhold a substantial portion of existing funding rather than new funding. Whether BEAD funding conditions can legally compel states to abandon their AI regulatory frameworks is a question that courts have not yet decided.
U.S. Secretary of Commerce
U.S. Attorney General
White House Special Advisor for AI and Crypto
Governor of California
Governor of Colorado
U.S. Representative (D-VA), Co-Chair of Congressional AI Caucus