FTC chair suggests AI developers should be liable for conduct of agents

US Federal Trade Commission Chairman Andrew Ferguson cautioned against portraying AI agents as autonomous entities capable of independent decision-making, arguing that the developers who instruct these agents should bear responsibility for any resulting harm. Speaking at the Reuters Momentum AI Austin event, Ferguson emphasized his intent to resist "anthropomorphizing" these tools, stating that when someone directs an agent to perform a task and the agent does it, "I don t think we would say, Oh, what do we do about the tool?"

The remarks indicate potential strategies for the Trump administration to adopt as instances of agentic AI testing leading to unauthorized access to sensitive data increase. While AI companies sometimes describe systems as functioning beyond human control, subsequent audits have revealed that the systems were merely executing the instructions they were given.

Ferguson stated that governments and industry leaders should examine whether existing oversight and cybersecurity measures are adequate to address these concerns.