FTC Chair Signals Liability Focus as OpenAI, Anthropic and METR Face Scrutiny Over AI Agent Incidents

U.S. Federal Trade Commission Chair Andrew Ferguson has pushed back on treating AI agents as independent actors, saying responsibility should remain with the people and companies directing them as scrutiny grows around recent frontier-model security incidents. The remarks land amid investigations and disclosures involving OpenAI, Anthropic and outside evaluator METR, but publicly available FTC materials do not show a newly announced agency probe naming OpenAI, Anthropic and METR together.