On July 1, 2026, the Federal Trade Commission (“FTC”) issued a proposed policy statement addressing what it describes as the “suppression of accuracy” in artificial intelligence (“AI”) systems and is seeking public comment through July 31, 2026. The proposal was issued pursuant to Executive Order 14365, Ensuring a National Policy Framework for Artificial Intelligence, which directed the FTC to explain how Section 5 of the FTC Act applies when AI developers alter model outputs in response to state law requirements.[1]
The proposed statement reflects the FTC’s continued focus on AI under Chairman Andrew Ferguson, but differs from earlier AI initiatives that largely focused on allegedly deceptive claims about AI capabilities.[2] Instead, the proposed statement focuses on whether AI systems are designed to pursue undisclosed objectives that differ from those users reasonably expect. According to the FTC, a company may violate Section 5 if it steers outputs toward undisclosed ideological or other objectives in a manner that conflicts with consumers’ expectations regarding the accuracy and objectivity of AI-generated responses, including where such changes are made to comply with state AI law requirements.
The FTC’s Deception Theory and its Implications for State AI Laws
The proposed policy statement builds upon the premise that AI companies have made both explicit and implicit representations that their systems are designed to provide accurate, objective, and useful outputs, subject to technological limitations. According to the FTC, consumers reasonably expect AI systems to pursue those objectives based on both companies’ marketing claims and the core value proposition that AI systems are tools for solving problems and providing information. The FTC therefore takes the position that a company may engage in deception if it intentionally designs its systems to pursue undisclosed objectives that differ from users’ reasonable expectations.
The FTC grounds this theory in Section 5 deception principles, arguing that affirmative misstatements, omissions, implied misrepresentations, and inadequate disclosures may all mislead consumers about the objectives an AI system is designed to pursue. The proposal identifies several examples of conduct that could raise concerns, including modifying outputs to advance ideological goals, changing outputs in response to political or public pressure, or altering outputs to avoid liability under certain state laws. The FTC specifically cites Colorado’s Artificial Intelligence Act as an example of a law that could incentivize AI developers to modify outputs in ways that depart from users’ expectations because it imposes potential liability on AI companies for discriminatory outcomes arising from their customers’ use of AI systems.[3] According to the FTC, if compliance with a state law causes a company to steer outputs away from users’ expected objectives without adequate disclosure, that conduct may constitute deception under Section 5.
Notably, the FTC distinguishes intentional output steering from AI “hallucinations,” stating that inaccuracies resulting from technological limitations generally do not raise the same concerns. The proposed statement also suggests that companies may mitigate potential liability through clear and conspicuous disclosures, and emphasizes that such disclosures must be sufficiently prominent to alter consumer expectations and cannot be buried in terms of service or fine print. Finally, the FTC suggests that state laws requiring conduct it views as deceptive may be preempted where they conflict with federal consumer protection law—a position that aligns with the Administration’s broader effort to establish a national AI policy framework and limit what it characterizes as a patchwork of state AI requirements.
Key Takeaways and Next Steps
The proposed policy statement provides insight into how the FTC may apply existing deception principles to the provision of AI systems going forward. In particular, the FTC indicates that it will focus on whether AI developers are sufficiently transparent about the objectives their systems are designed to pursue and whether those objectives align with users’ reasonable expectations. The proposal highlights the Administration’s scrutiny of state AI regulation and may foreshadow future disputes over the interaction between state AI laws and federal consumer protection principles.
The FTC is accepting public comments through July 31, 2026. Organizations interested in submitting comments should evaluate how the proposal may affect AI governance practices, model design decisions, disclosure strategies, and state law compliance obligations.
Covington will continue to monitor developments related to this proposed policy statement and related enforcement activity. If you have any questions concerning the material discussed in this blog post, please contact the members of our Advertising and Consumer Protection Investigations practice.
[1] Exec. Order No. 14,365, 90 Fed. Reg. 58,499 (Dec. 11, 2025).
[2] See, e.g., In re DoNotPay, Inc., FTC Docket No. C-4812, Decision and Order (Jan. 14, 2025) (addressing alleged deceptive claims regarding the capabilities of an AI-powered “robot lawyer”).
[3] Colo. Sen. Bill 26-189, 6-1-1707 (enacted May 14, 2026).