Draft Executive Order Seeks to Short-Circuit AI State Regulation

President Trump is preparing to sign an Executive Order that would seek to forestall state regulation of artificial intelligence (AI) by threatening federal lawsuits and the withholding of some federal funds. The draft, unsigned six-page Executive Order, “Eliminating State Law Obstruction of National AI Policy” (EO), the text of which has been circulating publicly since November 19, would declare it the policy of the Administration “to sustain and enhance America’s global AI dominance through a minimally burdensome, uniform national policy framework for AI.”

The White House is reportedly holding the EO in abeyance while Congress debates including a provision to preempt or otherwise attempt to block state-based AI regulation in the must-pass National Defense Authorization Act for Fiscal Year 2026 (NDAA).

The EO would not have the force of law nor be self-executing. Rather, it would require further action by federal agencies to implement and would likely trigger political and legal challenges by impacted states. Absent clear national legislation, the contested state and federal landscape will continue to impose significant shifting regulatory burdens on companies developing, integrating, and deploying AI systems.

Directing Agencies to Act Against State AI Laws

Specifically, the draft EO:

• Directs the U.S. Department of Justice to establish within 30 days an AI Litigation Task Force (Task Force) to sue states for enacting laws that, in the Administration’s view, unconstitutionally regulate interstate commerce or are preempted by existing federal regulations;
• Requires the Secretary of Commerce within 90 days to publish an evaluation of state AI laws that are “burdensome,” should be referred to the Task Force, or may require disclosures or reporting contrary to the First Amendment of the U.S. Constitution;
• Mandates the Secretary of Commerce within 90 days to issue a policy for agencies to condition certain federal broadband funding on whether those states enact or enforce AI laws that the Administration deems “burdensome”;
• Directs the Federal Communications Commission to begin a process to determine whether to adopt a federal reporting standard for AI models that would purport to preempt state laws;
• Directs the Federal Trade Commission to issue a policy on its power to prosecute unfair and deceptive trade practices to challenge state laws that “require alterations to the truthful outputs of AI models,” which the EO considers “DEI”—an undefined term presumably referring to the concept of diversity, equity, and inclusion; and
• Instructs White House officials to draft a legislative recommendation for a uniform federal regulatory framework for AI that would preempt state laws.

It particularly criticizes a recent California AI transparency law and Colorado’s AI Act for their allegedly burdensome reporting requirements and allegedly DEI-focused algorithmic discrimination provisions, respectively.

Curtailing State Lawmaking

The draft EO attempts by executive action what the U.S. Congress has failed to do through legislation—preempt or place a “moratorium” on state AI laws, which have proliferated in the past three years. State lawmakers have introduced hundreds of bills and adopted scores of them in state capitals across the country to protect consumers and children, limit AI use in certain circumstances, and impose transparency and reporting requirements on some AI developers.

In July 2025, the White House released “America’s AI Action Plan,” an extensive policy roadmap that exhorted the federal government to seek AI “dominance” by minimizing most regulations. The plan and related presidential orders led to a White House-issued request for information on federal, but not state, AI laws and policies that “unnecessarily hinder” AI development. The Commerce Department also requested input on a government-run American AI export program. And the Office of Management and Budget (OMB) has begun drafting guidelines barring the federal government from procuring AI systems that incorporate “ideological biases or social agendas,” including “diversity, equity, and inclusion.” (We expect OMB to issue those guidelines soon.)

In September, the White House touted the tech sector’s “transformative partnership with the Trump Administration” after the president hosted a widely publicized meeting with U.S. business leaders on “power[ing] American AI dominance.” And in mid-November, President Trump posted on social media his view that the Congress should pass a federal law to preempt state laws, criticizing in particular what he characterized as state efforts to “embed DEI into AI models.”

Attempts to include a preemption provision or a quasi-moratorium in the NDAA have thus-far sputtered in the face of Democratic opposition. Moreover, earlier attempts at legislation to impede state AI regulations have drawn opposition from many Republicans in Congress who support at least some state-based AI regulation.

Impact

While this EO would direct specific actions by federal agencies, its impact would remain to be seen, even if signed. For example, nowhere does the EO define what a “burdensome” regulation is, how to interpret “DEI,” and how agencies can determine what are “truthful outputs of AI.” Moreover, attempts to enforce these provisions—by bringing suits for violations of interstate commerce, withholding broadband funding, investigating state laws as unfair and deceptive, or preempting state laws based on an agency policy—will likely draw vigorous defenses or lawsuits from state attorneys general and impacted parties. Prominent Democrats and Republicans both criticized the draft EO as too closely aligned with industry and contrary to federalist principles, suggesting it will face stiff opposition.

Finally, federal preemption by executive decree is not a generally accepted practice under the U.S. Constitution and prevalent theories of separation of powers. Courts are usually “even more reluctant” to find state laws preempted based on mere regulations as opposed to statutes, and the U.S. Supreme Court has held recently that the anti-commandeering principles of the Tenth Amendment bar the federal government from prohibiting a state from legislating in a particular sector.

If this EO is nonetheless signed as written, state laws will likely remain in effect while the federal agencies proceed to implement the EO and even during the pendency of any federal government suits against states. Thus, the evolving, complex, and cross-cutting federal and state regulatory currents will continue, and companies that develop, integrate, and deploy AI systems should remain attuned to their shifting obligations.

International Complications

The Trump Administration’s approach to the “deregulation” of AI must also be placed in contrast to numerous laws outside the United States that directly impact those developing AI and looking to deploy it. For example, the EU AI Act, Data Act, and GDPR need to be considered in many use cases and scenarios, even by businesses that are U.S.-based. In addition, just this past week, there has been further upheaval with the European Commission announcing potential changes to these digital laws as part of its digital omnibus proposals to streamline rules on AI, cybersecurity, and data. AI regulation and compliance have, therefore, just become even more complex, quite aside from developments with this proposed EO.

***

Crowell & Moring will continue to monitor these fastmoving AI policy and legislative developments at the state, federal, and international level. For further information, please contact our team.