The ACLU has asked the Federal Trade Commission to withdraw its proposed policy statement on AI “accuracy suppression,” telling the agency the rule violates the First Amendment and recharacterizes state civil rights laws as statutes that force AI models to “produce false results.” The filing, reported by Inside AI Policy, reframes what the Commission has pitched as a consumer-protection measure into something closer to a preemption vehicle aimed at Colorado, California, and Illinois.

The statement, published in the Federal Register on July 7, 2026, cleared the Commission 2-0 and traces its authority to Executive Order 14365, signed December 11, 2025, and marketed under the banner of “Preventing Woke AI in the Federal Government.” Comments closed July 31. Roughly 40 had been logged on Regulations.gov as of the Federal Register notice.

The mechanism is worth naming plainly. The FTC proposes to use its Section 5 deception authority to treat state anti-discrimination requirements as impliedly preempted whenever they demand “alterations to the accurate outputs of AI models.” EO 14365 supplies the framing: a “minimally burdensome national policy framework for AI — not 50 discordant State ones.”

EFF, joined by Public Knowledge and Fight for the Future, filed parallel comments calling the proposal unconstitutional and warning it would install the FTC as “the arbiter of truth.” Tech Policy Press notes the statement “fails to argue meaningfully for preemption,” and flags that Chairman Ferguson himself once dissented from a related Section 5 application as “in direct conflict with the guarantees of the First Amendment.”

That last detail is the tell. The Commission is building a preemption doctrine on a theory its own chairman previously called unconstitutional. If finalized, litigation from state attorneys general is the obvious next stop.

Sources

Sources