Trump's AI Executive Order: Theater, Threats, and What Actually Holds Up in Court
President Donald Trump's new executive order offers bold talk on blocking state AI regulations. Legal experts say it reads more like pressure tactics than enforceable law - at least for now.
What the order says
The order declares a national policy to "dominate" AI with minimal regulation and directs federal agencies to push back on state AI laws. It tasks the Department of Commerce with producing, within 90 days, a list of state laws deemed "onerous" or in conflict with minimal federal oversight - including any that would "alter truthful outputs" or compel disclosures that allegedly violate First Amendment rights.
It instructs Attorney General Pam Bondi to create a litigation task force to challenge state laws under theories like federal preemption and the Interstate Commerce Clause. It also asks the FCC to consider a national disclosure standard that could preempt state rules and tells the FTC to issue a policy statement on when laws compelling alterations to "truthful outputs" might be deceptive.
The order threatens to withhold some federal broadband funds and unspecified grants from states with "onerous" AI rules, leaving agencies to decide what they can actually block.
Why legal experts call it "political theater"
"A lot of the order is really political theater," said Gowri Ramachandran of the Brennan Center for Justice, calling it "press-release-as-executive-order." The document is vague about which state laws it targets and outsources the heavy lifting to agencies that may lack clear authority.
Nicole Ozer of UC Law San Francisco noted the order is not law. Agencies can study, coordinate, and signal intent - but they can't nullify state statutes without a federal mandate to enforce. Congress hasn't passed major new consumer tech rules in decades, let alone a comprehensive AI statute.
The legal ceiling: preemption, commerce, and funding limits
- Preemption requires a statute. Agencies can't preempt state law by executive order alone. They need clear congressional authorization and valid rulemaking tied to that authority. Without that, preemption claims are thin.
- Dormant Commerce Clause is a stretch. Suits arguing that state AI laws unduly burden interstate commerce face an uphill path. Courts look for concrete burdens and extraterritorial reach; many AI safety and disclosure laws are written with in-state application. See overview: Dormant Commerce Clause.
- Spending powers are constrained. The administration can't withhold funds Congress ordered it to spend unless conditions are lawful, unambiguous, and related to the program - and not coercive. Prior attempts to claw back funds have often been blocked by courts. Reference: Spending Clause.
- First Amendment arguments cut both ways. Expect fights over compelled disclosures (commercial speech) and alleged requirements to "alter truthful outputs." The viability depends on whether laws are disclosure-only, how they're framed, and whether they meet standards for preventing deception or harm.
States aren't standing down
California and others have moved ahead with AI rules addressing safety, transparency, and self-harm risks. In New York, a bill to restrict deployment of high-risk models drew heavy industry pushback; Gov. Kathy Hochul moved to replace it with a lighter framework.
Momentum remains at the state level. The Senate voted 99-1 this summer against an effort to roll back state AI rules, and polling in California shows strong public support for tougher guardrails.
What's most likely to happen next
- Watch for agency signals, not immediate preemption. Expect lists, task forces, and policy statements first. Any serious preemption would require a final rule backed by statute - and would face litigation.
- Targeted lawsuits will test edge cases. The administration may pick specific state provisions to challenge (e.g., extraterritorial effects, compelled speech claims). Outcomes will be fact-specific.
- Funding threats will trigger suits. Attempts to block broadband or grant dollars will likely be enjoined if conditions weren't clearly authorized by Congress.
Practical guidance for legal teams
- Map state AI exposure now. Inventory applicable state AI statutes and bills (safety, transparency, testing, incidents, and model risks). Tie each to product features and deployment plans.
- Tune disclosures to survive scrutiny. If your state laws require risk, safety, or limitation disclosures, align with commercial speech standards and evidence of consumer protection. Keep records to defend against "compelled speech" challenges.
- Prep for dual compliance. Assume a period of coexisting state rules with federal "soft" guidance. Build policy modules that adapt by jurisdiction without rewriting core systems.
- Document safety controls. Where laws address self-harm mitigation or catastrophic risk, maintain a defensible record: evaluations, red-team findings, incident response, and prompt remediation timelines.
- Plan for litigation posture. For companies, evaluate potential Dormant Commerce Clause or First Amendment claims narrowly and strategically; success will vary by statute. For states, sharpen legislative findings and tailoring to local harms to withstand those challenges.
- Follow agency dockets. Track FCC and FTC actions hinted by the order; any attempt at preemptive standards or policy statements could shape compliance approaches even if not immediately enforceable.
Bottom line
The executive order tries to chill state AI laws with the threat of preemption and funding leverage. Without congressional action, its teeth are limited.
States will keep legislating. Companies should build compliance for a patchwork environment and be ready to litigate narrow issues - not bank on the order to sweep state rules off the table.
Your membership also unlocks:
