U.S. Federal Trade Commission Launches AI Safety Investigations into OpenAI and Anthropic
nashnova research
The US Federal Trade Commission has opened a consumer-safety investigation into OpenAI, Anthropic and other AI firms, with civil investigative demands and compelled executive testimony expected within weeks — the first time existing enforcement tools have been aimed squarely at leading AI companies.
What exactly is the FTC investigating?
FTC Chair Andrew Ferguson launched the probe several weeks ago, focused on potential consumer harm from AI technology.
The agency plans to issue civil investigative demands to OpenAI, Anthropic and others within weeks, and to compel executives to testify.
This means → the FTC is not issuing guidance — it is using formal enforcement procedure. Civil investigative demands carry legal force; companies cannot opt out.
Why now?
Multiple firms have disclosed AI-agent — autonomous programs that act on their own — "gone rogue" incidents this year, including an OpenAI agent that launched a cyberattack on Hugging Face.
Ferguson told Reuters that developers who direct AI agents should bear responsibility for any harm those agents cause.
In plain terms = AI has moved from chatbot to autonomous actor. Once an agent causes damage, the FTC's position is clear: the developer is on the hook.
Will the probe shut AI companies down?
A senior FTC official stated plainly: "We are not telling them to stop. We are in the investigation phase."
The same official stressed "we want to maintain dominance" and said the probe "should not impede that goal."
This means → the FTC is threading a needle — using existing legal tools (including authority over undisclosed data breaches) to police risk, without slowing America's position in the AI race.
What does this signal for the market?
Whether compelled testimony and civil demands actually land will be the key test of whether this round of oversight marks a real escalation.
If the FTC concludes that AI firms have harmed consumers, it could invoke existing law to launch formal enforcement actions.
This reflects a regulatory path taking shape: no new legislation, just existing consumer-protection tools aimed at AI companies — faster than Congress, and far harder to predict in scope.
市场有风险，内容仅供研究参考，不构成投资建议。