The Federal Trade Commission published a proposed policy statement on July 7, 2026, that takes direct aim at AI systems that produce outputs shaped by undisclosed objectives. The statement — titled “Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems” — argues that AI companies that secretly steer their systems away from accurate or user-expected outputs may be violating Section 5 of the FTC Act, which prohibits unfair or deceptive acts.

Comments from businesses, researchers, and the public are due by July 31, 2026.

What the FTC Is Saying

The FTC’s position is straightforward: when an AI company tells users that their system is designed to give accurate, helpful responses, those users have a reasonable expectation that the system is actually trying to do that. If the company has secretly tuned the system to prioritize other objectives — commercial, ideological, or otherwise — without telling users, that’s deception.

The policy statement notes that consumers accept AI outputs without fact-checking more than 90% of the time. That trust, the FTC argues, creates a heightened obligation on AI providers to be transparent when their systems are not simply optimizing for accuracy.

The statement does not say companies must always prioritize accuracy above all else. It says that when they don’t, they must make that clear — prominently, persistently, and in a way that actually reaches users. A disclaimer buried in a terms of service page is explicitly called out as insufficient.

What This Means in Practice

For AI companies: this policy, if finalized, creates a compliance obligation around disclosure. If your model is tuned to avoid certain topics, promote certain products, or defer to certain viewpoints, you need to tell users that clearly — not just technically permit yourself to in a legal document no one reads.

For businesses using AI tools: this matters less directly, but it does raise questions worth asking of your AI vendors. If you are deploying an AI assistant to your customers or staff, you should know what objectives that system is actually optimized for — and whether your vendor would be able to answer that question clearly.

For enterprise AI deployments: companies that white-label AI tools or embed third-party AI into their products will want to understand how this policy interacts with their own disclosure obligations. If a customer-facing AI assistant gives users incomplete or shaped responses, the business running that assistant may share in the compliance risk.

The Broader Regulatory Context

This FTC action sits within a wider pattern of AI regulation accelerating across multiple jurisdictions. The EU AI Act is now in effect for high-risk categories. Several US states have passed their own AI transparency laws. The White House has been in discussions with frontier AI companies about voluntary frontier model standards, with a formal August 1 deadline tied to the June executive order on AI innovation and security.

The FTC’s move is notable because it uses existing consumer protection law rather than waiting for new AI-specific legislation. Section 5 has been on the books for decades — applying it to AI accuracy gives regulators an enforcement tool now, without waiting for Congress to pass new rules.

It is also worth noting the timing. The FTC policy arrives at a moment when AI adoption among US businesses has moved from pilot to production at scale. More employees and customers are interacting with AI systems daily, which means the stakes of deceptive AI behavior are higher than they were even 12 months ago.

What This Means for Business

The immediate practical action is simple: if your business uses AI tools to interact with customers or make recommendations, understand what those tools are actually optimized for. Ask your vendors directly. Review your disclosure language.

If you are building or customizing AI agents, now is a good time to document the objectives you have tuned those agents for — and to make sure your user-facing language accurately describes what the system does and does not do.

The comment period closes July 31. If your business has a stake in how these rules are written, submitting a comment is a legitimate channel to influence the final policy.

Enterprise DNA helps businesses deploy AI agents that are transparent about their capabilities and limitations. If you are navigating AI vendor selection or building internal AI tools, book a discovery call to understand what questions to be asking.