FTC chair: agent developers bear liability, audit logs are the shield

On September 25, 2026, FTC Chairman Andrew Ferguson told an audience at the Reuters Momentum AI event in Austin that liability for the conduct of AI agents lies with the developers who build them, not with the agents themselves. “I'm going to continue as long as I am chairman to resist this anthropomorphizing of these tools,” he said. His reasoning, as reported, was that if someone tells a tool to do something and the tool does it, the scrutiny belongs on the instruction and on the toolmaker rather than on the tool.

In the same appearance, Ferguson emphasized the evidentiary weight of audit records. According to the DEV Community report published on October 4, 2026, he said that when a supposedly “rogue” agent is being reviewed, the audit trail is what should be examined. The report's headline put the position bluntly: when an agent causes trouble, the developer is on the hook, and the log is the liability shield.

That DEV Community article is the only coverage of the remarks available here. It does not indicate whether the FTC has opened any proceeding or investigation, and it names no company or product that might be affected. There is also no mention of whether the chairman's comments connect to any existing enforcement matter, guidance, or rulemaking.

As of the article's publication, the position Ferguson stated remained a public remark delivered at a conference. The material gives no indication of whether it will be followed by formal FTC documents, policy, or further statements.