Executive Order 14365, "Ensuring a National Policy Framework for Artificial Intelligence," is a Trump administration executive order signed on December 11, 2025 and published in the Federal Register at 90 FR 58499 on December 16, 2025. It establishes a federal policy of preempting state AI laws that the administration deems to hinder innovation, create regulatory fragmentation, or impose ideological bias on AI models, and it directs the creation of an AI Litigation Task Force, an evaluation of state AI laws, federal funding restrictions, and a recommendation for preemptive federal legislation.
Stated purpose
The order frames its objective as a national policy framework intended to prevent a patchwork of state AI regulation. It characterizes the Colorado AI Act's ban on "algorithmic discrimination" as potentially "forc[ing] AI models to produce false results," and it identifies state laws that, in the administration's view, hinder innovation, fragment regulation, or impose ideological bias as the targets of the federal response.
Key provisions
AI Litigation Task Force (Sec. 3). The Attorney General must establish a task force within 30 days to challenge state AI laws that unconstitutionally regulate interstate commerce, are preempted by existing federal regulations, or are otherwise unlawful.
Evaluation of state AI laws (Sec. 4). The Secretary of Commerce must publish within 90 days an evaluation identifying state laws that "require AI models to alter their truthful outputs," laws that compel disclosures violating the First Amendment, onerous laws conflicting with federal pro-innovation policy, and laws that should be referred to the Litigation Task Force.
Restrictions on state funding (Sec. 5). States with "onerous AI laws" may be made ineligible for federal broadband (BEAD) funding. Agencies must assess whether discretionary grants can be conditioned on states not enacting or not enforcing identified AI laws.
Federal preemption standard (Secs. 6–7). The FCC is directed to initiate a proceeding on whether to adopt a federal reporting and disclosure standard that preempts conflicting state laws. The FTC is directed to issue a policy statement explaining when state laws requiring AI to alter "truthful outputs" are preempted by the FTC Act's prohibition on deceptive practices.
Legislative recommendation (Sec. 8). The order directs preparation of legislation establishing a uniform federal AI policy framework that preempts conflicting state AI laws, with carve-outs for child safety, AI compute and datacenter infrastructure, state government procurement, and other topics to be determined.
Relation to state AI regulation
The order positions the federal government against state-led AI regulation. It explicitly names the Colorado AI Act as an example of problematic state law, and its provisions reach state transparency laws including California SB 53 and the NY RAISE Act. The funding restrictions in Section 5 create a mechanism to condition federal money on states declining to enact or enforce identified AI laws.
The order represents a direction of preemption aimed at preventing state regulation rather than creating new substantive requirements. It can be contrasted with other approaches to AI regulation documented across related laws: frontier-transparency disclosure requirements (SB 53, RAISE Act); product-liability regimes (AI LEAD Act); anti-discrimination requirements (CO AI Act); and antitrust enforcement through case law.
| Approach | Example | Direction | ||
|---|---|---|---|---|
| Frontier transparency | [[california-sb-53\ | SB 53]], [[new-york-raise-act\ | RAISE Act]] | More disclosure |
| Product liability | [[ai-lead-act\ | AI LEAD Act]] | More liability | |
| Anti-discrimination | [[colorado-ai-act\ | CO AI Act]] | More equity | |
| Antitrust | Case law | More competition | ||
| Federal preemption | EO 14365 | Less state regulation |
The dynamic the order sets in motion — the federal government reclaiming regulatory space that states had occupied — relates to the account of state-federal division in Techno-Federalism (Wu), in which states filled a federal vacuum on AI regulation. The order reverses that direction by seeking to roll back state action.
Points of tension with other sources
The order's approach conflicts with several positions documented elsewhere. It opposes Dario Amodei's argument in The Adolescence of Technology that transparency legislation such as SB 53 is the right starting point, framing such laws as potentially unconstitutional. Its premise that AI models should produce "truthful outputs" stands in tension with the Emergent Misalignment research, which finds that models cannot be reliably controlled to produce only truthful outputs. The ceiling-preemption design of EO 14365, which would bar stronger state protections, also conflicts with the floor-preemption design of the AI LEAD Act, which preserves the ability of states to adopt stronger protections.
Provenance
This page summarizes Executive Order 14365 (signed December 11, 2025; published at 90 FR 58499 on December 16, 2025) from the primary text (Source: Raw Sources/Executive Order 14365 - Ensuring a National Policy Framework for Artificial Intelligence.md).