FTC and AI Output Steering

The Federal Trade Commission’s proposed policy statement applies Section 5 of the FTC Act to AI systems intentionally designed to steer outputs away from objectives consumers reasonably expect, such as accuracy or objectivity.

It distinguishes deliberate design choices from ordinary model errors and considers whether prominent disclosures can meaningfully change users’ expectations. The proposal also raises questions about state AI laws: Colorado’s Artificial Intelligence Act is a focal example of how compliance requirements might affect outputs, while the FTC’s position raises potential federal-state conflict and preemption issues. The policy subject centers on how consumer-protection principles apply to AI behavior and the claims made about it.