President Trump issued an Executive Order (EO)¹ that aims to take steps towards a uniform national framework for artificial intelligence (AI), primarily by attempting to curb the effect of state-level AI laws.  The Order may yield legal challenges, which would set the stage for a sustained test of the Administration’s preemption efforts in the absence of actual federal statutory enactment. 

The President’s announcement follows Congress’s bipartisan decision to exclude from the final text of the National Defense Authorization Act (NDAA) a provision that would have preempted most state AI laws.  That decision marked the second high-profile failed attempt to attach an AI preemption measure to major legislation, following a similar omission of a 10-year (then 5-year) moratorium on state AI laws in the tax-and-spending bill ultimately signed into law earlier this year. 

The Order seeks to displace what it describes as a “patchwork” of “50 discordant State” regulatory regimes that, in the Administration’s view, “thwarts” the innovation necessary for United States to “win[] the AI race.”  It singles out Colorado’s AI Act as an example of harms posed by state AI laws, arguing that its ban on “algorithmic discrimination” compels entities to “embed ideological bias” within models—potentially forcing AI to produce “false results” to avoid “‘differential treatment or impact’ on protected groups.”

To replace this “patchwork,” EO announces that the Trump administration will “act with the Congress to ensure” a “minimally burdensome national standard” that preempts laws in “conflict with the policy set forth in this order,” and among other things, protects children and respects copyrights.   Until that national standard is established, however, the Order asserts an immediate need to check “onerous and excessive” state regulations and, to that end, directs executive branch components to pursue four initiatives:

1. Department of Justice (DOJ) Litigation Task Force: The EO directs the DOJ to establish an “AI Litigation Task Force” charged with bringing challenges to designated state AI laws.
2. Conditions on Federal Funding: The EO instructs federal agencies to assess whether discretionary funding programs can be conditioned on a state’s agreement to forgo the types of AI regulation that are disfavored under the order. It further instructs the Secretary of Commerce to withhold certain funds for broadband access from states that maintain “onerous” AI laws identified by the Secretary and the AI Litigation Task Force.
3. Regulatory Action: The EO directs the FCC to initiate a proceeding to adopt a “Federal reporting and disclosure standard” for AI that would preempt conflicting state laws. It also directs the FTC to issue a policy statement identifying the circumstances under which state laws requiring changes to “truthful outputs” are preempted by the FTC Act’s prohibition on deceptive practices.
4. Legislative Recommendation: The EO requires senior Administration officials to develop a legislative proposal for a uniform federal regulatory framework for AI that would preempt any state laws deemed inconsistent with the order’s stated policy. The EO states the legislation shall not propose preempting state laws regarding child safety protections, AI compute or data center infrastructure, or state government AI procurement.

Below, we describe these initiatives in greater detail and identify notable legal dimensions of the EO’s approach.

EO Initiatives and Potential Legal Challenges

The “AI Litigation Task Force” (Section 3)

The EO directs the DOJ to establish, within 30 days, an “AI Litigation Task Force” whose “sole responsibility shall be to challenge state AI laws.”  The Task Force will target, ultimately through civil litigation brought by DOJ, state laws that the Attorney General determines “unconstitutionally regulate interstate commerce,” are “preempted by existing Federal regulations,” or are “otherwise unlawful.”

Whether the AI Litigation Task Force has real impact on companies such as AI model developers and on states that have passed or are considering passing AI legislation depends primarily on its ability to generate successful legal challenges.  The Task Force will have to navigate substantial legal complexity in its effort to do so.  As a foundational principle, for a state law to be preempted by federal law, Congress generally must have legislated in the space—and, as noted, Congress has not enacted comprehensive legislation regulating AI.  That will leave the question for the Task Force and, if DOJ does bring lawsuits, for courts whether other federal legislation should be deemed to have preemptive effect with respect to particular state-level AI laws.

In addition, the EO’s direction that the Task Force should mount litigation challenges to state laws that “unconstitutionally regulate interstate commerce” appears premised on the Dormant Commerce Clause, which generally prohibits states from enacting laws that discriminate against, or unduly burden, interstate commerce.  The Task Force is likely to contend that state AI laws impose an excessive burden on interstate commerce by subjecting national providers to incompatible obligations and thus, in practice, forcing them to design their products—which are borderless and thus inherently difficult to geofence—around the most restrictive of these state standards.²

That theory, if pursued in court, will almost certainly meet opposition from affected states.  They will likely point, in particular, to the skepticism that the Supreme Court has recently expressed that extraterritorial economic effects alone render a law unconstitutional.³  States also are likely to emphasize that the Court has expressed doubt about attempts to rely on the Dormant Commerce Clause to invalidate state laws that neither discriminate against out-of-state entities nor reflect a state’s intent to protect its own industries at the expense of those of other states.⁴  Additionally, affected states are likely to raise threshold standing questions, as Dormant Commerce Clause challenges have traditionally been brought by private entities (and occasionally states) able to show concrete economic injury, rather than by the Executive Branch itself.⁵

Conditions on Federal Funding (Section 5)

The EO also seeks to leverage federal dollars to discourage state AI regulation.  Section 5(b) of the EO generally directs agencies to determine whether they may condition other discretionary grants on a state’s agreement not to enact or enforce conflicting AI laws.  Section 5(a) more specifically directs the Secretary of Commerce to specify that states with AI laws identified as undermining AI dominance would be ineligible for funds under the Broadband Equity, Access, and Deployment (BEAD) Program, a broadband access program.  This provision asserts that “a fragmented State regulatory landscape for AI threatens to undermine BEAD-funded deployments” and the “growth of AI applications reliant on high-speed networks.”  The EO also instructs federal agencies to assess whether they can condition discretionary grants on states entering into a “binding agreement … not to enforce” conflicting AI laws.

This approach to conditioning federal funds is also likely to be resisted by states if they in fact are deprived of federal funding on the grounds set out in the EO.  They will, in particular, likely point to indications from the Supreme Court that conditions attached to federal grants must generally relate to the federal interest in the particular program.⁶  States are likely to argue that AI safety and algorithmic-governance rules—aimed at model training, bias mitigation, or “catastrophic risk”—bear no meaningful relationship to BEAD’s core purpose, which is the physical deployment and operation of broadband infrastructure.  Courts will then need to consider whether the asserted connection between broadband deployment and AI model governance is too “attenuated” or that these grant conditions either exceed the President’s authority or—if the agency’s actions are challenged—are arbitrary and capricious for want of a reasoned explanation.⁷

Agency Guidance (Sections 6 & 7)

The EO seeks to block state AI laws indirectly through regulatory action by the FCC and FTC.  Section 6 would require the FCC to “initiate a proceeding to determine whether to adopt a Federal reporting and disclosure standard for AI models that preempts conflicting State laws” within 90 days.  Section 7, in turn, would direct the FTC—again within 90 days—to issue a policy statement that would “explain the circumstances under which State laws that require alterations to the truthful outputs of AI models” (for example, bias‑mitigation or content‑adjustment mandates) are preempted by the FTC Act’s prohibition on deceptive practices.  The FTC’s policy statement must also describe the agency’s view on how the FTC Act’s prohibition on unfair and deceptive acts or practices would apply to AI models.

Assuming that the FTC and FCC conclude that state AI requirements can be displaced by their governing statutes or other federal law, implementation of these directives too will likely face legal resistance from affected states and/or companies—raising, most fundamentally, the question of whether federal law supersedes state law, which is generally contingent on federal legislation or regulation, not executive prerogative.  Federal agencies can, in some circumstances, adopt regulations with preemptive effect, but only when they are acting within the scope of authority that Congress has delegated by statute.⁸

Whether the Communications Act or the FTC Act empowers the agencies to displace state AI laws will likely be tested by any affected states, which long have been regulating unfair and deceptive practices alongside federal agencies.  Moreover, states are likely to challenge DOJ’s likely argument that existing statutes authorize the executive branch to direct the FCC and FTC to preempt state AI and consumer-protection laws.

Conclusion

The EO advances the Administration’s goal of a national AI framework—without federal statutory changes—by invoking preemption, the Dormant Commerce Clause, and the Spending Clause, among other issues.  While the EO does not have immediate and direct effect on states or companies, its implementation could carry significant consequences, especially if novel executive branch actions survive anticipated litigation challenges by affected states.

WilmerHale boasts a wide range of attorneys with expertise and experience in AI-related matters.  That includes lawyers in the firm’s regulatory, litigation, transactional, and intellectual property practices.  The firm works regularly with AI model developers, AI model testers, and AI model deployers, including counseling them on the latest developments in federal, state, and international AI legislation and regulation.  We are well placed to advise affected parties on the newly issued EO and its anticipated implementation in the months to come.