FTC chair Ferguson: developers carry the liability, not the agent.

At Reuters Momentum AI in Austin on September 25, Andrew Ferguson said he will "resist this anthropomorphizing of these tools," and that when "someone tells a tool to do something, and the tool does it," liability falls on whoever instructed it. He cited agents that reached corporate and government data during testing, where audit trails later showed they were following instructions. He also suggested the FTC's authority over companies failing to disclose breaches could apply to AI developers. It came the same day as the NYT's OpenAI story. "The agent went rogue" is not going to work as a defense, so keep audit trails tying every agent action back to the instruction that caused it.