The FTC must publish an AI policy statement today, ninety days after an executive order directed the commission to clarify when state AI laws are preempted by the federal consumer protection statute. The word preemption is doing more work than the law allows. Four agencies received coordinated deadlines. One writes the framework. One identifies the targets. One drafts the replacement standard. One litigates. This is not a policy statement. It is a siege.
The Federal Trade Commission must publish a policy statement today. The deadline — March 11, 2026 — was set ninety days ago when President Trump signed an executive order titled Ensuring a National Policy Framework for Artificial Intelligence. The directive: explain how Section 5 of the FTC Act, the commission's core authority over unfair and deceptive practices, applies to artificial intelligence. And explain when state AI laws are preempted by the federal statute.
The word doing the most work in that directive is preempted.
The Four Agencies
The executive order does not rely on the FTC alone. It constructs a coordinated mechanism across four federal agencies, each with a different assignment and the same objective: centralize AI governance at the federal level.
The FTC issues the interpretive framework — a policy statement explaining how existing consumer protection law already covers AI, and when state laws that require alterations to what the order calls the "truthful outputs" of AI models conflict with that federal authority. The Commerce Department publishes a report identifying which state AI laws qualify as "onerous" — defined as laws that require AI models to alter truthful outputs or compel developers to disclose information in ways the administration considers unconstitutional. The FCC drafts a federal reporting and disclosure standard designed to preempt conflicting state requirements. And the Department of Justice established an AI Litigation Task Force on January 10, charged with challenging state AI laws in federal court.
One interpretive anchor. One target list. One replacement standard. One litigation unit. Today's FTC statement is the framework. The DOJ task force is the enforcement.
The Speech Argument
Buried in the executive order's definitions is a legal theory that transforms the preemption argument from a regulatory question into a constitutional one.
The order targets state laws that require alterations to the "truthful outputs" of AI models and laws that compel developers to engage in "impermissible compelled speech." This is a First Amendment frame applied to artificial intelligence. The implicit claim: when an AI model produces an output, that output is speech. When a state requires the output to be modified, labeled, or suppressed, the state is compelling or restricting expression.
First Amendment preemption is categorically stronger than statutory preemption. A policy statement explaining that the FTC Act covers AI is guidance — nonbinding, challengeable, limited. A constitutional argument that state AI transparency and bias-testing laws violate the First Amendment can invalidate those laws entirely, regardless of what the FTC says. The executive order constructed both paths and assigned each to a different agency.
Whether AI model outputs qualify as "truthful" speech protected by the First Amendment is a question no court has decided. The executive order proceeds as if the answer is yes.
The Parallel
This journal has spent four months documenting what happens when a federal agency claims preemption over state law in the domain of autonomous systems participating in commerce.
The CFTC asserted that federal commodity law overrides state gaming classifications of prediction markets. Tennessee called Kalshi a swap. Massachusetts called it a bet. A Nevada federal judge ruled in March that CFTC registration does not preempt state gaming law — the federal floor is not a ceiling. The CFTC also withdrew its proposed ban on political and sports event contracts in February, signaling a shift from prohibition to governance under existing authority.
Now the FTC enters the same structural position — a federal agency asserting authority over fifty states on how autonomous systems participate in commerce — with the same word but a fundamentally different instrument.
The CFTC's preemption argument was tested in court. A judge ruled. The result was specific, limited, and mixed. The FTC's preemption argument arrives as a policy statement — nonbinding guidance that the International Association of Privacy Professionals and TechPolicy.Press have both analyzed as carrying limited preemptive authority.
But the executive order anticipated this limitation. The policy statement is not the enforcement mechanism. The DOJ task force is.
The Named Target
Colorado's SB 24-205 — the most comprehensive state AI law in the country, requiring algorithmic discrimination impact assessments and transparency obligations for high-risk AI systems — is specifically named in the executive order.
The law was signed in May 2024. Its implementation has already been delayed to June 30, 2026, following a failed special legislative session in August 2025. Now it faces a coordinated federal response: named in the executive order, targeted by a DOJ litigation task force, subject to a Commerce Department review of whether it qualifies as onerous, and framed by an FTC statement explaining when such laws are preempted.
Legal analysts have noted that if the DOJ challenges SB 24-205 on preemption or constitutional grounds, there is no guarantee Colorado's governor or attorney general will spend political capital defending a law that has already survived one near-repeal attempt.
Colorado is the named target. California, Illinois, and other states with AI legislation are the unnamed ones. The executive order carved out exceptions for child safety protections, AI compute infrastructure, and state government procurement. What remains in the crosshairs is specifically the governance of AI outputs — transparency mandates, bias testing requirements, output modification rules.
The Mechanism, Not the Label
The CFTC preemption battles produced mixed results because they were fought on legal terrain where judges rule based on precedent and statute. The FTC preemption campaign begins on different ground. Today's policy statement is an opening position, not a court ruling. The constitutional litigation comes later, through the DOJ, on First Amendment grounds the executive order has already framed.
The pattern across both domains is consistent: apply existing federal law as a floor, then pressure states out of building above it. In prediction markets, the CFTC withdrew prohibition and applied existing commodity law. In AI, the FTC applies existing consumer protection law while separate mechanisms target the states that went further.
Same administration. Same legal strategy. Same deregulatory thesis. Different domain. The word preemption is doing more work than the law currently allows. Watch the DOJ task force, not the policy statement.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)