The Federal Trade Commission's AI accuracy proposal is now in its public-comment phase, giving companies, researchers and consumers until July 31, 2026, to weigh in on how the agency should treat claims about artificial intelligence systems.

The proposal, published in the Federal Register on July 7, asks how the FTC Act's ban on deceptive acts or practices should apply when companies market AI tools that users expect to answer accurately, neutrally or according to stated objectives.

The practical point is simple: this is not a final AI rule. It is a proposed policy statement, and the current step is a public record-building process before the agency decides what guidance to finalize or how to use it in future enforcement.

What the FTC is asking

The FTC says AI systems are increasingly used for work, education, research and everyday decision-making. Its proposed statement focuses on whether a company can mislead users if it advertises an AI product as accurate or user-directed while quietly steering outputs toward a different objective.

The agency frames the issue under Section 5 of the FTC Act, the long-running consumer protection law that bars unfair or deceptive acts or practices. In the notice, the FTC asks for comments on the proposed statement and on how the policy should address companies that market AI systems to the public.

The FTC's public comment page lists the matter as Document ID FTC-2026-0859-0001. The agency's announcement says the Commission authorized the Federal Register notice by a 2-0 vote, and that comments will be posted after processing.

Why it matters

For AI users, the key question is not whether a chatbot, search assistant or workplace tool is always correct. The issue is whether a company clearly tells users what the tool is optimized to do and whether its outputs match the claims made in marketing, onboarding screens or sales materials.

For developers and businesses buying AI products, the proposal is a reminder to separate product claims from product hopes. If a tool is tuned for brand safety, legal caution, speed, engagement, political neutrality, customer support escalation or another goal, users may need clear disclosures before they rely on the output.

The proposal also matters because it lands while states, schools and employers are writing their own AI rules. A federal policy statement would not replace every state law, but it could shape how the FTC views deceptive marketing claims when AI systems produce answers that users reasonably believe are objective or complete.

What to check now

Consumers should treat the comment window as a signal to read AI product claims carefully. Look for clear statements about accuracy limits, data freshness, source citations, safety filters, personalization settings and whether outputs are generated to satisfy the user's request or another policy goal.

Organizations using AI tools should review customer-facing language, procurement materials and internal training. The safer question is not just whether the model performs well in testing, but whether the company can support the promises it makes about how the tool behaves.

The next date to watch is July 31, 2026, when comments are due. After that, the FTC can review the record and decide whether to revise, finalize or abandon the policy statement.