FTC Proposes AI Accuracy Rule to Stop Deception

The FTC published a policy statement in the Federal Register on July 7, 2026 proposing to apply the agency’s prohibition on deceptive acts or practices under Section 5 of the FTC Act to companies that market AI systems. The filing states that steering or masking the true capabilities of an AI to meet external demands, such as state rules or misrepresenting performance to win customers, could amount to deception under the agency’s standards.

The document notes a real world incentive structure that could tempt firms to tilt AI outputs to satisfy outside pressures. Colorado’s recently revised Artificial Intelligence Act is cited as a cautionary example of how state requirements might collide with consumer expectations, prompting some sellers to adjust outputs in ways that obscure accuracy. The White House’s January 2026 study, Artificial Intelligence and the Great Divergence, is referenced to frame a broader tension: balancing innovation, consumer protection, and the risk of misalignment between capability claims and real world performance. The policy statement is careful to distinguish legitimate compliance with laws from deceptive distortion, suggesting that the line lies in consumer “reasonable expectations” about what an AI can or cannot do.

For compliance officers and product leaders, the implications are immediate. The filing states that marketing claims about AI accuracy, reliability, or the system’s ability to perform tasks should be grounded in demonstrable data and transparent about limitations. In practice, that means codifying internal governance around claims, building verifiable benchmarks, and establishing disclosure controls before any product goes to market. The risk is not only a misleading launch pitch, but downstream harm when customers rely on incorrect outputs in critical decisions. The FTC’s approach would make deceptive marketing of AI a formal target for enforcement, not a peripheral concern.

Enforcement mechanisms are a core part of the proposal. The policy statement would extend the FTC Act’s deception prohibition to AI marketing, giving the agency a familiar toolkit to address violations. While the notice does not spell out every remedy, observers expect the FTC to rely on its existing authority to pursue corrective actions, injunctions, and other measures against companies that falsely tout AI performance or conceal material inaccuracies. The move signals a sharpening regulatory lens on how AI capabilities are presented to buyers, users, and enterprises, not just how they are developed.

Industry watchers identify four practical implications to watch in the coming months. First, compliance teams should revamp external communications, ensuring every claim about AI accuracy comes with verifiable data and clear caveats. Second, product and engineering leaders must implement robust testing, logging, and documentation so claims can be independently verified rather than asserted from hype. Third, legal and risk officers will need to map multi jurisdictional exposure, given the reference to state acts such as Colorado’s AI law, and prepare cross border governance. Fourth, executives should internalize a higher enforcement risk premium in go to market plans, balancing speed to scale with the need to avoid deceptive marketing that could trigger FTC action.

The policy is a proposal, not a final rule, and it points to ongoing engagement with stakeholders. The FTC’s action reflects a broader push to align AI performance claims with consumer protections, a trend that could reshape how AI vendors market, price, and govern their products in the months ahead. The filing shows the commission is serious about policing accuracy claims and steering, with enforcement under its longstanding Section 5 authority likely to define practical consequences for noncompliant firms.