FTC Opens Probe Into AI-Agent Risks
The FTC has opened an investigation into OpenAI, Anthropic and METR focused on potential consumer risks from increasingly autonomous AI agents.
AI-agent safety is moving from research discussions into regulatory scrutiny as autonomous systems gain access to more tools and user data.
What happened
The US Federal Trade Commission opened an investigation into OpenAI, Anthropic and AI-safety research organisation METR focused on potential consumer risks from increasingly autonomous AI agents.
The agency is expected to seek information from the organisations and may request executive testimony.
Opening an investigation does not establish wrongdoing by any of the organisations involved.
Why it matters
Agents can take actions rather than merely generate text.
That means failures can involve account access, purchases, communications or changes to external systems.
The relevant safety question therefore extends beyond whether a model produces an incorrect answer.
Regulators increasingly need to understand how companies control permissions, monitor actions and prevent systems from exceeding what users intended.
The bigger picture
AI regulation is shifting from model-level questions toward deployment architecture.
As agents become embedded in real products, consumer protection may depend on audit logs, identity controls, human approval and limits on autonomous action.
That creates both compliance pressure and a market opportunity for security companies building the infrastructure required to make agents observable and controllable.