FTC proposes policy statement targeting deceptive claims about AI accuracy

The Federal Trade Commission has proposed a policy statement describing how it may apply the federal ban on deceptive acts or practices to companies that market artificial-intelligence systems while suppressing, obscuring or misrepresenting information about those systems’ accuracy.

The proposal, posted July 6, 2026, does not itself establish a new AI-safety rule or announce that a particular company has violated the law. Instead, it signals the commission’s possible approach to marketing claims about AI performance, including what companies disclose—or do not disclose—about accuracy.

What the proposal says

The policy statement concerns Section 5 of the FTC Act, the statutory provision identified by the commission as the legal basis for addressing deceptive acts or practices. The proposal focuses on companies marketing AI systems and on claims or representations about whether those systems are accurate.

In particular, the proposal addresses the suppression of accuracy information in AI marketing. That means the FTC is examining not only affirmative claims about how well a system performs, but also whether relevant information about accuracy is hidden, obscured or withheld in a way that could make marketing deceptive.

The proposal does not establish that any particular AI system is inaccurate. Nor does the existence of the proposal amount to a finding that a named company has engaged in deceptive conduct. Those questions would depend on the facts of individual matters and any future action by the commission.

Why it matters for AI companies

The proposal could affect how AI developers and marketers test, describe and document their products. Companies promoting an AI system may face greater pressure to substantiate performance claims and to consider whether information about accuracy is presented clearly enough for consumers and other customers to evaluate the system.

The practical significance is therefore less about an immediate change to what every AI company must do and more about the enforcement risk the FTC is identifying. The agency’s position could influence product testing, advertising and disclosure practices across the sector, particularly where marketing emphasizes performance without adequately explaining accuracy-related information.

The issue can arise across different types of AI products because the proposal is framed around the marketing of AI systems and the treatment of accuracy information, rather than around one named technology or one industry. The approved FTC record does not provide a list of covered products or specify a single accuracy standard.

Where the proposal stands

The FTC identifies the document as “FTC-2026-0859-0013.” It was also published under Federal Register number 2026-13628. Those identifiers distinguish the proposal in the commission’s administrative and public-record systems.

A policy statement generally explains enforcement principles. It is not itself a new statute or regulation, and the packet does not identify the proposal as a binding requirement. The proposal therefore represents the FTC’s stated direction and potential enforcement framework, not a final rule that automatically changes the legal obligations of AI developers or marketers.

The FTC’s broader enforcement materials list 2026 activity involving privacy, AI and consumer data. That page includes related actions involving AI-powered marketing, student data and sensitive location data. Separately, the Justice Department’s Antitrust Division lists ongoing federal technology and digital-market matters, including actions involving Google and other digital-market issues. Those materials provide broader federal enforcement context, but they do not establish that those matters are part of this proposed policy statement.

What happens next

The source record confirms that the FTC has posted the proposed policy statement, but it does not provide a comment deadline or say that the commission has adopted a final policy statement. The next confirmed step is therefore the proposal’s public consideration and any later action the FTC may take.

Until the commission takes further action, the central question for companies is how they support and present claims about AI accuracy. For consumers, the proposal is a signal that the FTC may scrutinize not only plainly false performance claims, but also marketing practices that leave out or conceal information needed to assess those claims.

Sources

• Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems, Federal Trade Commission
• Privacy and Security Enforcement, Federal Trade Commission
• Antitrust Division, U.S. Department of Justice

Look for updates to this story

Related News

• Senate examines the consumer cost of AI-driven surveillance pricing
• Federal prosecutors seek victims in alleged $100 million nonprofit trust fraud case
• Publishers and Authors Sue Google Over Gemini Training Data
• State Watchdog Finds Gaps in USAID Program Transfer