← All news

Policy

FTC seeks comment on proposed AI accuracy and hidden-steering policy

An FTC draft argues that AI companies could deceive consumers when they hide changes made to answers for goals users did not request or expect. The proposal leaves basic questions about accuracy, safety limits, and federal authority unresolved.

reviewedUpdated Jul 15, 2026, 6:59 AM UTC
Original source

Federal Trade Commission

Read the original source

What happened

The Federal Trade Commission issued a proposed policy statement on July 1 and opened public comments through July 31. The commission approved the notice by a 2-0 vote. This is a proposal, not a final rule or a finding that any company broke the law.

The draft says an AI company could violate the FTC Act when it secretly changes outputs to serve an 'ideological objective' that conflicts with reasonable consumer expectations. It distinguishes those choices from ordinary mistakes caused by technical limits.

The FTC draft also argues that federal law may override conflicting state AI laws, and it singles out Colorado's revised AI Act. Independent legal analyses describe that position as unsettled. One says courts will likely decide the question.

Why it matters

A final policy could affect how AI companies describe accuracy, disclose other goals, and explain safety or legal limits on answers. It could also sharpen disputes over whether federal consumer-protection law can displace state AI rules.

What remains unclear

Related claims

Sources

  1. FTC Seeks Public Comment on Policy Statement Addressing AI AccuracyPrimary source - Federal Trade Commission - regulator - Jul 1, 2026

    Used for: The proposal date, 2-0 commission vote, July 31 comment deadline, stated consumer-protection theory, and the fact that the policy is not final.

    Open source

  2. Federal Trade Commission's Proposed Policy Statement Concerning The Suppression Of Accuracy In Artificial Intelligence SystemsPrimary source - Federal Trade Commission - regulator - Jul 1, 2026

    Used for: The canonical proposed policy document and its framing around undisclosed objectives, consumer expectations, and possible conflict with state AI laws.

    Open source

  3. A view from DC: Consumers deserve the truth about AI systemsInternational Association of Privacy Professionals - news - Jul 10, 2026

    Used for: Independent analysis of the proposal's novel legal theory and the unresolved problem of defining the truth and accuracy consumers should expect from generative AI.

    Open source

  4. FTC Proposes Policy Statement on AI Accuracy and Ideological Manipulation of AI OutputsTroutman Pepper Locke - legal - Jul 2, 2026

    Used for: Independent legal analysis of the FTC's deception theory, disclosure standard, treatment of ordinary model mistakes, and specific preemption claim involving Colorado's revised AI Act.

    Open source

  5. What the Supreme Court's decision means for the FTCAxios - news - Jul 2, 2026

    Used for: Independent reporting that the FTC proposal asserts federal preemption only when state law conflicts with a federal regulatory scheme.

    Open source