Federal Trade Commission Chairman Andrew Ferguson signaled Friday that AI developers could bear legal responsibility for harmful actions carried out by agents operating under their instructions.
Ferguson rejected the idea that AI agents should automatically be treated as independent actors with intentions separate from the people or companies directing them.
His comments provide an important indication of how the FTC could approach liability as businesses deploy increasingly autonomous software capable of interacting with websites, databases, consumers and commercial systems.
Rather than creating an entirely separate legal category for autonomous agents, Ferguson indicated that the United States could use existing legal tools.
He also suggested that FTC authority covering companies that fail to properly disclose data breaches could potentially apply to AI developers.
The issue has become more urgent as incidents involving AI agents accessing corporate and government systems draw scrutiny from regulators and cybersecurity researchers.
Ferguson’s position could have significant implications for companies developing agentic systems.
If regulators apply existing consumer-protection, privacy and cybersecurity rules directly to AI developers and deployers, companies may need stronger audit trails, permission controls and monitoring systems capable of showing exactly what instructions an agent received and what actions it subsequently performed.
The FTC is separately examining personalized pricing and how companies may use detailed consumer information when determining prices.
For AI companies, the immediate policy signal is that increasing software autonomy may not necessarily provide a liability shield.
The next question is how aggressively the FTC and other regulators apply existing enforcement authority as agentic AI becomes more widely deployed.
