In December 2025, the White House escalated its effort to rein in the fast-expanding landscape of state artificial intelligence regulation. A new executive order (EO) frames state-by-state rules as a growing barrier to a coherent national AI strategy, and signals that the federal government is willing to use both litigation and funding leverage to reduce what it calls an increasingly fragmented compliance environment.
The move comes amid a surge of state activity: the White House says more than 1,000 AI-related bills have been introduced in state legislatures. At the same time, major states are still pressing forward with new laws and sector-specific restrictions, setting up a near-term clash over federal preemption, interstate commerce, and where the line sits between protecting residents and regulating innovation.
1) The Executive Order’s Core Aim: A “Minimally Burdensome” National Framework
The Dec. 11, 2025 executive order makes its intent unusually explicit: it directs the federal government to pursue a “minimally burdensome” national AI framework and criticizes the “patchwork of 50 different regulatory regimes.” The message is that regulatory fragmentation is itself a policy problem, one that, in the administration’s view, threatens the pace of AI deployment and U.S. competitiveness.
The White House fact sheet puts the argument in economic and geopolitical terms. It quotes President Trump advocating “one Federal Standard” rather than a state patchwork, warning that overregulation could allow “China” to catch up. This framing treats regulatory uniformity not merely as administrative convenience, but as a strategic imperative.
That posture matters because executive orders are often indirect, nudging agencies rather than drawing bright lines. Here, the order is designed to produce concrete outputs on a short clock, including legal challenges and a roadmap for federal legislation that would preempt conflicting state rules while preserving certain carve-outs.
2) From Guidance to Courtroom: The DOJ “AI Litigation Task Force”
One of the order’s sharpest tools is the directive to the Department of Justice to form an “AI Litigation Task Force” within 30 days. The EO says this unit’s “sole responsibility” is to challenge state AI laws deemed inconsistent with the order’s policy, explicitly including arguments grounded in the Commerce Clause and federal preemption.
This is a notable escalation from the typical federal approach of issuing model guidance or offering cooperative federalism incentives. By creating a dedicated litigation shop, the administration signals it expects disputes to be resolved through injunctions, declaratory judgments, and appellate precedent, potentially accelerating the timeline on which a preemption doctrine for AI could emerge.
In practice, the task force could prioritize cases where a state law effectively regulates conduct beyond the state’s borders, burdens interstate AI services, or conflicts with federal agency mandates. Even if some suits fail, the mere prospect of sustained federal litigation can reshape how states draft bills and how companies assess compliance risk.
3) Commerce Department’s 90-Day Review: Identifying “Onerous” State AI Laws
The EO also directs the Department of Commerce to publish, within 90 days, an “evaluation of existing State AI laws.” The evaluation must identify “onerous” laws that conflict with the EO’s policy and flag candidates for litigation referral, creating a pipeline from policy review to courtroom action.
That pipeline is important because it turns what could have been a broad rhetorical critique of state regulation into a structured enforcement workflow. In effect, Commerce becomes the triage function: mapping the state AI landscape, selecting likely targets, and generating an evidentiary record that can be leveraged in lawsuits or negotiations.
The federal government is also positioning itself to influence legislative design before laws take effect. If Commerce identifies a proposed or newly enacted requirement as “onerous,” states may face immediate pressure to amend, delay enforcement, or risk becoming a test case for federal challenge.
4) The “Truthful Outputs” Focus and First Amendment Fault Lines
The EO’s minimum criteria for Commerce’s evaluation specifically targets state laws that “require AI models to alter their truthful outputs.” It also targets laws that compel disclosures or reports that may violate the First Amendment. These are not generic administrative concerns; they are constitutional fault lines.
The “truthful outputs” language suggests the administration is preparing arguments that certain AI regulations function as compelled speech or content-based restrictions, areas where courts apply heightened scrutiny. While states may argue they are regulating product safety, consumer protection, or professional standards, the EO frames some requirements as speech mandates that distort truthful information.
This focus could reshape the drafting of state AI rules. Legislators may attempt to shift from output-based mandates (e.g., “must say X”) toward process-based standards (e.g., testing, risk management, audit trails). But even process mandates can raise constitutional questions if they indirectly force changes to outputs or editorial decisions embedded in model behavior.
5) FTC Preemption Signaling: A Policy Statement on Deception and State Mandates
Alongside DOJ and Commerce, the EO assigns a distinct role to the Federal Trade Commission. Within 90 days, the FTC Chair is directed to issue a policy statement explaining when state laws requiring alteration of “truthful outputs” are preempted by the FTC Act’s ban on deceptive practices.
This is an unusual legal pivot: the FTC Act is generally associated with consumer protection and unfair or deceptive acts, not with shielding firms from state requirements. But the administration appears to be setting up an argument that a state-mandated alteration of truthful information could itself create deception, thereby colliding with federal consumer-protection objectives.
If the FTC issues an aggressive preemption-oriented statement, it may influence courts even if it is not binding. Companies challenging state laws often cite agency interpretations to support conflict-preemption theories, arguing that state rules frustrate federal purposes or impose inconsistent obligations that increase the risk of deception.
6) Funding Leverage: BEAD “Non-Deployment” Funds and Broader Grant Conditioning
Beyond litigation, the administration is also using fiscal tools. The White House fact sheet says the order directs Commerce to withhold certain BEAD-related “non-deployment” funds from states with identified “onerous” AI laws, and encourages other agencies to consider similar grant conditioning.
After publication of the EO in the Federal Register on Dec. 16, 2025, the text underscored that states with “onerous AI laws” can be made “ineligible for non-deployment funds” under BEAD “to the maximum extent allowed by Federal law.” The same publication contemplates conditioning other discretionary federal grants on states not enacting conflicting AI laws, or agreeing not to enforce them during the funding period.
This approach can be powerful because it changes the bargaining dynamics. States may have strong policy reasons to regulate AI, especially in sensitive areas like healthcare, education, elections, or child protection, but may hesitate if the price is losing important broadband or technology-planning funds. Expect significant debate over whether such conditions are sufficiently related to the grants and whether they cross constitutional lines governing federal spending conditions.
7) States Keep Moving: New York’s RAISE Act and “AI Companion” Rules
Despite the EO, states have continued legislating. The Wall Street Journal reported that New York enacted the “RAISE Act,” which requires safety plans for certain large AI developers and mandates breach reporting within 72 hours, with an effective date of Jan. 1, 2027. That timing gives industry and regulators a runway, but it also creates a clear future enforcement target if federal agencies decide the law is “onerous” or conflicting.
Reuters also reported that New York and California enacted early state rules for “AI companions,” highlighting how states are beginning to draw lines around emotionally responsive systems and other high-salience use cases. These are precisely the kinds of niche categories where state experimentation tends to happen first, before federal legislation catches up.
The result is a practical stress test of the EO’s preemption posture. If prominent states keep passing targeted laws, especially those that touch on speech, disclosures, or safety-plan requirements, the DOJ task force and Commerce evaluation process may quickly face decisions about whether to litigate, negotiate, or tolerate limited state variation.
8) What Comes Next: Congress, Preemption, and Sector-by-Sector Uncertainty
The administration has acknowledged that durable uniformity likely requires legislation. Reuters quoted White House adviser Sriram Krishnan saying the administration will work with Congress on a single AI framework, pointing to “over 1,000” state-level regulations and emphasizing that legislation will ultimately be needed.
The EO also directs preparation of a legislative recommendation for a uniform federal AI framework that would preempt conflicting state laws, while carving out categories the order says should not be preempted, such as certain child-safety protections. Those carve-outs will be politically decisive: they shape whether federal law is seen as a ceiling (limiting states) or a floor (allowing states to go further in specific areas).
In the meantime, sectors with heavy state involvement, like healthcare, are watching closely. The American College of Radiology noted the EO may affect healthcare AI rules via litigation risk and funding leverage, even before Congress acts. Law-firm analyses, including Skadden’s, have also outlined potential constitutional and preemption theories that could become central if the DOJ begins challenging state statutes at scale.
The White House move to preempt state AI laws is not a single instrument but a coordinated strategy: identify “onerous” state requirements, pressure states through federal funding, clarify agency preemption positions, and litigate selected laws to establish court precedent. Together, these tools aim to replace rapid state experimentation with a more uniform national framework, on terms defined by the executive branch while Congress is urged to codify the end state.
Whether this strategy succeeds depends on legal limits, political will, and how states respond. If large states continue enacting targeted AI rules, especially in emerging categories like AI companions or in developer obligations like safety plans and breach reporting, federal courts may soon become the primary arena for deciding how far states can go before federal policy, the First Amendment, and interstate-commerce principles draw a boundary.
