FTC Enforcement Against Deceptive AI Claims

Section 5 of the FTC Act prohibits unfair or deceptive acts or practices in or affecting commerce. The FTC has applied that authority to AI claims, data use, and products. Operation AI Comply was a law-enforcement sweep announced in September 2024 with actions involving, among others, an AI fake-review tool, an “AI lawyer” service, and business-opportunity schemes promising income from AI storefronts. The Commission’s complaints allege violations; allegations are not final findings unless resolved through an order or judgment.

A claim can be deceptive when it is likely to mislead reasonable consumers about a material fact. Claims that a tool is more accurate, safer, autonomous, profitable, or equivalent to a human professional should have a reasonable basis before they are made. The strength of evidence should match the claim. A benchmark on a narrow task does not substantiate a broad promise of professional-quality advice or guaranteed earnings. Disclosures do not necessarily cure an overall misleading impression.

The FTC also enforces privacy and security promises. If a company tells customers their confidential data will not train models, it should honor that commitment, including through vendors and workarounds. The Commission has said that remedies in prior privacy cases can include deletion of products, models, or algorithms developed using unlawfully obtained data; whether such relief applies depends on the facts and legal order. It is inaccurate to imply that every AI case leads to model deletion.

An AI product team should inventory public claims, preserve substantiation, test performance on representative use cases, and disclose meaningful limits. Keep evidence contemporaneous and connected to the exact version and user population. Treat complaints, refund patterns, and failure reports as signals to review claims and controls. When an FTC matter is announced, distinguish a complaint, proposed order, final administrative order, and court judgment.