The Trump administration’s recent executive order, titled “Ensuring a National Policy Framework for Artificial Intelligence,” asserts broad federal authority to preempt state laws regulating AI in order to establish a “minimally burdensome national standard” for AI regulation.1 Although the Executive Order signals the Trump administration’s intent to use legal challenges to preempt state laws, it remains far from certain that the push will succeed. Aggressive efforts by the administration to preempt state laws could provoke considerable resistance from states and other stakeholders, and the legal theories the administration is likely to use may face significant obstacles in court. This article explores the legal backdrop for the Executive Order, its key provisions, and key legal questions raised by the legal theories that the administration is likely to advance.
I. The Existing Federal and State Regulatory Frameworks for AI
Since inauguration, the Trump administration has expressed that fostering AI innovation is a core administration priority. Through a series of executive proclamations, it has emphasized the need to eliminate what it views as “burdensome” regulations and to encourage AI adoption by federal agencies, with the stated goal of positioning U.S. companies to lead the global AI race. In January 2025, the administration issued an Executive Order that rolled back the Biden administration’s minimum safeguards for so-called high-risk AI—meaning AI that is used in ways that could threaten public safety or rights—which the current administration characterized as “barriers” to innovation.2 The Trump administration continued its deregulatory push through “America’s AI Action Plan,” issued in July 2025, which directed agencies to “remove barriers” to AI development and repeal regulations seen as hindering innovation.3 Simultaneously with these policy efforts, the administration has invested in expanding federal AI usage through programs like the Genesis Mission, an effort intended to develop an integrated AI platform to leverage federal scientific datasets and accelerate scientific research and technological development.4
These policy pronouncements have not been matched by any significant federal legislation or regulation of AI. In the absence of a federal framework, states have taken the lead in shaping the AI regulatory landscape. In 2025 alone, all 50 states, Washington D.C., Puerto Rico, and the U.S. Virgin Islands introduced AI-related legislation, with 38 states adopting or enacting such laws.5 State AI laws address a range of requirements across multiple sectors, including consumer protection, health care, employment, and financial services. Among other things, these laws require disclosure when AI is used; risk assessments and safeguards to mitigate bias and unlawful discrimination; patient consent to AI use in certain clinical settings; meaningful human oversight of specified decisions, such as health care prior authorization determinations; and governance controls over the use and collection of AI-related data.
For example, Colorado enacted the Consumer Protections in Interactions with Artificial Intelligence Systems Act, which requires deployers of high-risk AI systems to use reasonable care to protect consumers from known or reasonably foreseeable risks of “algorithmic discrimination.”6 Texas enacted the Responsible Artificial Intelligence Governance Act, which requires disclosure to consumers when they are interacting with AI and prohibits development or deployment of AI systems with the intent to unlawfully discriminate.7 Numerous other states have enacted or are considering similar comprehensive AI regulations, consumer protection measures, and requirements for human oversight of AI in health care decisions.
Over the past year, the federal government has considered two significant legislative attempts to preempt state regulations but has declined to enact either bill. In connection with the “One Big Beautiful Bill Act,” Congress considered a provision that would have preempted state AI regulations for a period of 10 years. After significant media scrutiny and bipartisan opposition, the Senate voted 99-1 to strip the state preemption provision from the bill.8 Congress then declined to enact a similar state AI preemption moratorium through the 2025 National Defense Authorization Act.9
II. The Latest AI Executive Order
The Executive Order takes up the baton from these failed legislative efforts.10 In keeping with the administration’s prior executive statements and actions relating to AI, the Executive Order declares that it is “the policy of the United States to sustain and enhance the United States’ global AI dominance through a minimally burdensome national policy framework for AI.”11 It states that the administration “must act with the Congress to ensure that there is a minimally burdensome national standard.”12 In the meantime, the Executive Order asserts that the administration will take action to “check the most onerous and excessive laws emerging from the States.”13
The Executive Order asserts that state-by-state regulation has created “a patchwork of 50 different regulatory regimes” that makes compliance more challenging, particularly for start-ups, and that state laws “sometimes impermissibly regulate beyond State borders, impinging on interstate commerce.”14 The Executive Order also asserts that state laws are “increasingly responsible for requiring entities to embed ideological bias within models” by targeting “algorithmic discrimination,” singling out Colorado’s AI Act as “forc[ing] AI models to produce false results in order to avoid ‘a differential treatment or impact’ on protected groups.”15
The Executive Order states that the administration seeks to minimize the impact of state AI regulation through both litigation challenging state laws the administration views as inconsistent with its AI policy and regulatory steps to limit state regulations. Key provisions of the Executive Order include:
- Federal AI Litigation Task Force: Directs the United States Attorney General to create an “AI Litigation Task Force” with the “sole responsibility” of challenging state AI laws, including on grounds that such laws are unconstitutional regulations of interstate commerce, are preempted by existing federal law, or “are otherwise unlawful in the Attorney General’s judgment.”16
- Identifying State Laws for Scrutiny: Requires the Secretary of Commerce to publish an evaluation of “onerous” state AI laws, particularly those that in the Secretary’s view require AI models to “alter their truthful outputs” or mandate disclosure of information in violation of the First Amendment.17
- Potential Federal Reporting and Disclosure Standard: Instructs the Federal Communications Commission (FCC) to “initiate a proceeding” to determine whether to adopt federal AI reporting and disclosure standards that would supersede conflicting state laws.18
- Defining Unfair and Deceptive Acts or Practices: Directs the Federal Trade Commission (FTC) to issue a policy statement clarifying the circumstances under which state laws that mandate “alterations to the truthful output of AI models” are preempted by federal law prohibiting unfair and deceptive acts or practices.19
- Federal Funding Conditions: Requires the Secretary of Commerce to issue a policy notice stating when states with restrictive AI laws may be deemed ineligible for certain Broadband Equity Access and Deployment (BEAD) Program funds. Other federal agencies must also assess whether they can condition discretionary grants under other programs on states refraining from enacting or enforcing conflicting AI laws.20
- Legislative Proposal: Charges the Special Advisor for AI and Crypto and the Assistant to the President for Science and Technology with drafting legislation that would broadly preempt state AI laws, with exceptions for child safety, data center infrastructure, and state government procurement.21
III. A Primer on Preemption
The Executive Order signals a broad effort by the administration to preempt state AI law, but the Executive Order does not itself have immediate effect on states and stakeholders. Understanding what effect the Executive Order will have on state AI laws requires an overview of the doctrine of federal preemption.
A. What Is Preemption?
Under the Supremacy Clause of the Constitution, where a conflict arises between federal and state law, federal law can displace state law through the doctrine of preemption. Two main categories of preemption exist: express preemption and implied preemption.22 Express preemption applies where a federal statute or regulation explicitly states that its provisions supersede state law.23 Implied preemption applies where the preemptive effect of federal law is implicit in the structure and purpose of the law.24 The theory of preemption that could most likely be implicated by the Executive Order is implied conflict preemption, under which state law is preempted to the extent that it poses an obstacle to the accomplishment of federal objectives.
Critically, preemption is not automatic. The Executive Order, standing on its own, lacks preemptive force, as it is not a statute enacted by Congress nor a regulation enacted pursuant to congressional authorization.25 Moreover, even where applicable, federal preemption is not self-executing. As explained further below, parties can assert preemption in the context of lawsuits, either through an affirmative suit brought by the federal government or, under certain circumstances, by regulated parties, or as a defense to state enforcement efforts. However, unless and until courts invalidate state laws on preemption grounds, regulated parties must continue to comply with state AI regulations.
B. How Would Preemption Get Raised?
The Executive Order indicates that the federal government will use administrative levers to induce states to repeal or forgo enforcement of their own laws—including by restricting federal funding or by investigating companies that comply with certain state AI laws. However, legal challenges to state AI laws are likely to be a key mechanism for establishing the federal preemption of state laws deemed inconsistent with the administration’s deregulatory stance.
Generally, there are four main paths for raising preemption in court. In the first three, another party would initially raise the preemption argument, but the federal government could appear in court to support the party’s preemption argument. In the fourth, the federal government itself could directly bring suit.
First, a private party could raise preemption as a defense to an enforcement action brought by a state for a failure to comply with a state AI regulation, arguing that because federal law preempts the state law, the state law is void.
Second, a private party could sue a state official before the law is enforced to seek an injunction against its enforcement. Pre-enforcement suits against states brought by private parties are generally barred by state sovereign immunity under the Eleventh Amendment. However, the Ex Parte Young doctrine provides an exception for suits seeking to enjoin a current or future violation of the federal Constitution by a state officer if the state officer has actual authority to enforce the challenged law and there is a “credible threat” of such enforcement action against the plaintiff.26 Because some state AI regulations impose penalties for noncompliance, such as fines, parties may be able to bring pre-enforcement challenges to those AI laws.
Third, a state aligned with the Executive Order’s deregulatory objectives could challenge the AI laws of another state as preempted by federal law or an unconstitutional regulation of interstate commerce. This suit would not be barred by state sovereign immunity, but the state bringing the suit would need to establish Article III standing to sue in federal court, including identifying a direct concrete injury that the state AI regulation has on the state itself or, potentially, on the interests of its citizens.27
In any of these first three scenarios, the federal government would have the ability to appear in the litigation to support preemption. Federal law gives the Department of Justice broad authority to appear in any state or federal court “to attend to any . . . interest of the United States.”28 The federal government routinely appears in proceedings to weigh in on preemption questions by filing statements of interest and amicus briefs.
Fourth, the United States could also directly sue a state or state official to enjoin current or imminent enforcement of a state law on the grounds that it violates federal law. For example, in recent years, the federal government brought suits under the Supremacy Clause against Texas and Iowa for immigration laws that conflicted with federal immigration laws,29 against Missouri for a law declaring federal gun regulations “invalid” within the state,30 and against the New Mexico Supreme Court for a rule of professional conduct that would limit the circumstances under which prosecutors can subpoena a lawyer in a criminal proceeding for information about past or present clients.31
Such a suit would not be barred by state sovereign immunity,32 though the federal government, like any other litigant, must have Article III standing to bring suit in federal court. Traditionally, courts have been deferential when assessing the government’s standing to challenge state laws on Supremacy Clause grounds, in light of the government’s sovereign interest in enforcing its laws. Courts may scrutinize the government’s standing closely in this context, however, given the Executive Order’s acknowledgement that there is no current national regulatory standard for AI.
IV. Potential Headwinds to Federal Preemption
The federal government’s efforts to preempt state AI laws are likely to engender substantial pushback from states and other stakeholders. As noted above, the Executive Order is neither a statute nor a regulation, and it does not itself have the force of the law. As such, it is insufficient on its own to preempt state laws. Moreover, as the Executive Order itself acknowledges, there is no existing federal regulatory framework, which complicates the administration’s contention that state AI laws conflict with federal law. Tellingly, Congress has twice rejected efforts to enact preemption provisions, which could suggest that Congress does not intend to foreclose state regulation of AI.
In addition, the legal theories highlighted by the Executive Order as potential bases for federal challenges to state laws are potentially vulnerable to legal challenge. While the Executive Order does not discuss these theories in detail, the contours of some of these theories, and potential state responses to them, are visible.
We briefly touch upon some of the preemption theories relied upon by the Executive Order, below.
Dormant Commerce Clause:
The Executive Order identifies the Dormant Commerce Clause as a key theory to challenge state law. Under the Dormant Commerce Clause doctrine, states may not enact laws that unduly restrict interstate commerce. Any legal challenge on this ground is likely to posit that a particular state’s AI law unconstitutionally burdens or regulates interstate commerce. However, the Supreme Court recognized in a recent decision that state regulations are not invalid under the Dormant Commerce Clause merely because they impose effects beyond state borders.33 By contrast, laws that discriminate against out-of-state economic interests in favor of in-state interests are more susceptible to successful challenge under the Dormant Commerce Clause.
Courts applying this framework in an analogous context have rejected challenges to state laws regulating the content that internet service providers may offer, concluding that a state law that requires altering internet content within that state does not constitute an attempt to regulate out-of-state conduct, even if the practical effect of the law is not confined to the state’s borders.34 Likewise, a state sued by the federal government regarding its AI law may argue that the challenged law does not attempt to regulate conduct beyond the state’s borders, but instead, requires only that AI companies modify the content that they offer within that particular state. This argument may be stronger in the context of AI models, which may be more easily geofenced than content on the internet. Moreover, most state AI regulations do not expressly favor in-state AI developers over out-of-state AI developers, which indicates that these laws are likely to be less susceptible to challenge on Dormant Commerce Clause grounds.
FTC Act:
The Executive Order also highlights Section 5 of the FTC Act (15 U.S.C. § 45) as a potential ground for preemption. The Executive Order directs the Chairman of the FTC to issue a policy statement explaining the circumstances under which state laws that “require alterations to truthful outputs of AI” may be preempted under Section 5.35 The Executive Order’s focus on laws that require alterations to AI outputs likely relates to state laws the administration views as “embed[ding] ideological bias within models,” including Colorado’s law regulating algorithmic discrimination.36
How the administration intends to use Section 5 of the FTC Act is not apparent from the Executive Order. Section 5 prohibits unfair and deceptive acts or practices in commerce and authorizes the FTC to investigate and prosecute deceptive business conduct. Typical examples of enforcement actions brought under Section 5 involve false advertising, unfair methods of competition, violations of consumers’ privacy and data security rights, and telemarketing violations. States may levy several objections to any legal challenges relying on Section 5 as a basis for preemption. States are likely to contest whether Section 5 has preemptive effect, as Section 5 is not traditionally viewed as displacing state law. States generally enforce their own consumer protection laws alongside Section 5, and in fact, the FTC regularly collaborates with states regarding enforcement of Section 5. Furthermore, states will likely argue Section 5 is only enforced through civil penalties and thus cannot be used to invalidate state law. In addition, states may also raise a Major Questions Doctrine argument, arguing that using Section 5 to preempt state AI regulations would represent a significant expansion of the FTC’s authority regarding an issue of significant political and economic importance without clear congressional authorization.
Further, states may contest whether any particular state regulation constitutes an unfair or deceptive practice, and how the phrase “alterations to truthful outputs” should be interpreted in the context of AI models is uncertain. States may argue the Executive Order is directing the FTC to act beyond its authority of regulating business practices to govern outputs of AI models. For similar reasons, federal challenges to such laws may implicate First Amendment concerns to the extent the FTC’s preemption efforts attempt to target specific viewpoints.
Additional Potential Theories:
The Executive Order signals the possibility of additional legal grounds for preemption but leaves these theories largely unexplained.
The Executive Order calls for the FCC to “initiate a proceeding to determine whether to adopt a Federal reporting and disclosure standard for AI models” that would preempt state law.37 The FCC regulates interstate and international communications and is responsible for establishing an “efficient, Nation-wide, and world-wide” telecommunications service.38 Congress has given the FCC limited express preemption authority via Section 253 of the Communications Act, under which the FCC can preempt state law “to the extent necessary” if it determines that it “prohibit[s] or ha[s] the effect of prohibiting the ability of any entity to provide any interstate or intrastate telecommunications service.”39 However, the Executive Order does not explain how any future reporting standard that the FCC might promulgate would conflict with and preempt state law. Nor does it explain how Section 253, which applies to “telecommunications services,” might extend to AI. Stakeholders should monitor the development of any proposed FCC reporting standard to assess what state AI laws may be implicated.
The Executive Order also references preemption by “existing Federal regulations,” but does not specify which regulations may serve as a basis for such preemption.40 Similarly, the Executive Order also suggests that state laws may be challenged as “otherwise unlawful in the Attorney General’s judgment,” but provides no further guidance regarding this purported basis. Regardless, the Attorney General’s judgment is not, in and of itself, a basis for preemption, which must be grounded in duly enacted federal legislation or regulations issued pursuant to Congress’s authorization.41
V. Looking Ahead
The Executive Order envisions a uniform national AI policy guarded from contrary state regulation through the power of preemption. But this vision is not a legal reality. Although Congress has the authority to preempt state AI laws through legislation, it has thus far declined to do so. The Executive Order directs federal agencies to utilize existing federal laws and forthcoming regulations to challenge state laws that conflict with the administration’s deregulatory posture toward AI, but those efforts may face stiff resistance in court. The viability of the administration’s legal theories is uncertain, particularly in light of the Executive Order’s acknowledgement of the absence of a federal regulatory framework and Congress’s rejection of recent efforts to preempt state AI laws. Accordingly, it remains unclear how far the administration’s preemption arguments will sweep.
In the near term, stakeholders should closely monitor the litigation and policy landscape. The Executive Order is likely to result in the issuance of significant policies by federal agencies designed to preempt conflicting state laws, including a policy statement by the FTC on Section 5 of the FTC Act, a policy notice by the Department of Commerce on BEAD Program funding, potential regulatory efforts by the FCC, and potential conditions on federal grant funding.
In addition, stakeholders should expect a spate of litigation on preemption issues. Notably, the Department of Justice recently created a new Enforcement and Affirmative Litigation Branch within the Civil Division that is tasked with “filing lawsuits against states, municipalities, and private entities that interfere with or obstruct federal policies,” reflecting the administration’s priority of challenging state and local laws that it believes violate the Supremacy Clause.42 The administration is likely to utilize this new branch and its AI Litigation Task Force to pursue deregulatory efforts through lawsuits directly challenging state laws. In addition, the Executive Order may embolden private litigants or sympathetic states to initiate litigation against state AI laws, in the hopes that the federal government would appear in the litigation to support those efforts. Finally, the Executive Order could breathe a second wind into efforts by Congress to pass uniform national AI legislation that would preempt conflicting state laws or, alternatively, to clarify when state laws are not preempted by federal law.
Ropes & Gray continues to monitor developments related to state and federal AI regulations, including through its Health AI Atlas and Standing Orders, Local Rules, and Decisions on the Use of AI tracker.
