The FTC’s AI inquiry is approaching the point where companies must supply evidence

The Federal Trade Commission is close to sending civil investigative demands to frontier AI companies, a senior agency official told Semafor on October 8. Questions concern companies’ statements about dangers and consumer harm. The demands had not been confirmed as served.

The distinction matters because the inquiry was publicised unusually early, alongside a nonbinding industry safety accord. An announced investigation signals concern; compulsory information requests can begin building an evidentiary record.

There is also a dispute about process. Two people told Semafor that career staff learned of the inquiry through the press. The senior official rejected that account and said Bureau of Consumer Protection lawyers were drafting the demands. Former FTC chair William Kovacic described advance disclosure as unusual but not inherently improper.

Analysis: the useful next milestone is evidence collection, followed by a clear account of what the agency believes was misleading or harmful. Neither an investigation nor a demand establishes wrongdoing. But obtaining internal material would make this a more concrete test of the laboratories’ assurances than another voluntary declaration.