FTC AI accuracy docket draws 358 comments

The FTC’s AI accuracy docket received 358 comments, with 252 posted publicly, before its July 31 window closed, according to the live federal docket. The Commission’s 2–0 proposal says undisclosed ideological steering could violate Section 5 and argues conflicting state rules may be preempted. For AI vendors already facing consumer-protection liability over capability claims, the operator move is to preserve evaluation records and document intentional output controls before the FTC turns this proposal into final policy.