The Federal Trade Commission opened a public-comment proceeding over how Section 5 of the FTC Act might apply when companies market artificial-intelligence systems as objective or accurate but allegedly shape outputs for undisclosed ideological purposes. The agency announced the proposed policy statement July 1 and set a July 31 comment deadline.
The proposal treats the issue primarily as one of advertising and consumer expectations. Section 5 bars unfair or deceptive acts or practices. In the FTC's account, an AI seller could create a deception problem if its express claims, demonstrations or surrounding marketing lead users reasonably to expect objective or accurate results while undisclosed interventions push the system in another direction.
A second part addresses conflicts between federal and state approaches. The proposal singles out Colorado's Artificial Intelligence Act and argues that a state requirement could be impliedly preempted if it conflicts with a federal regulatory scheme. That is the commission's proposed legal position, not a court ruling.
The commission voted 2-0 to authorize publication of the notice. The docket was posted under document FTC-2026-0859-0001, and the public-comment page identified the matter as a proposal. The comment period has now closed; as of August 8, the FTC's AI topic page still labels the statement proposed rather than final.
For developers and sellers, the practical question is less whether an AI system can be perfectly neutral—a concept the notice does not define—and more whether the product behaves consistently with what the company tells customers. Claims about reliability, suitability and objectivity could become evidence in an enforcement inquiry if the policy is finalized and later applied.
The proceeding itself neither changes an AI model nor imposes a penalty. Any final statement, enforcement case or accepted preemption theory would require later action. The official development is a completed request for comment on how existing consumer-protection law should be interpreted in this setting.