On 7 July 2026, the Federal Trade Commission published proposed policy statement 2026-13628 in the Federal Register under the title Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems. The statement identifies a category of conduct it calls suppression of accuracy: deliberate modification of an AI model's outputs to advance undisclosed ideological objectives, causing the system to produce answers differing from what it would otherwise generate. The statement is at the proposed stage; the public comment period closes 31 July 2026.
The proposed statement invokes Section 5 of the Federal Trade Commission Act, 15 U.S.C. § 45, prohibiting unfair or deceptive acts or practices in or affecting commerce. Under the deception prong, an AI provider who represents its system as objective or accurate while secretly configuring it to steer outputs away from accurate answers makes a materially false representation likely to deceive a reasonable consumer. The statement further applies the implied preemption doctrine: a state statute that effectively requires an AI provider to suppress accurate outputs conflicts with Section 5's consumer protection purpose and is, to that extent, displaced by federal law.
AI system providers and deployers serving US consumers face two immediate considerations. First, any representation about the objectivity or accuracy of an AI system in terms of service, product documentation, or marketing materials must reflect the system's actual configuration; undisclosed output steering that contradicts those representations creates Section 5 exposure. Second, businesses that have built compliance measures around Colorado's Artificial Intelligence Act or analogous state statutes should assess whether those measures require suppressing accurate outputs and, if so, how the FTC's preemption argument affects their risk analysis.
The proposed statement does not create an enforceable rule; it articulates the Commission's enforcement posture and invites comment. The implied preemption claim applies only where a state statute effectively requires suppression of accurate outputs, a factual and legal determination requiring case-by-case analysis of the specific state law and system configuration at issue. Businesses should monitor whether the final policy statement modifies the preemption framing and whether Colorado or other states respond with legislative adjustments.
Licentium advises clients on AI regulatory compliance across US federal and state requirements. Work we undertake includes FTC Section 5 risk assessments for AI products and services, AI product documentation and disclosure review, US state AI legislation mapping, and cross-border AI governance strategy.