OpenAI, Anthropic Face New FTC Probe Over AI Agent Risks

The Federal Trade Commission (FTC) said Wednesday it has launched a sweeping investigation into major artificial intelligence developers, including Anthropic and OpenAI, to examine whether their products could put consumers at risk. The move marks the agency’s first U.S. enforcement effort focused on so-called AI agents and follows a string of incidents that drew public attention in July.

The commission intends to send compulsory information requests and seek testimony from leaders at prominent AI labs, naming Anthropic, OpenAI and research group METR. The New York Post published the initial report.

Anthropic and OpenAI have turned to METR for outside reviews tied to security events involving agent-style systems, the official said. 

Separately, President Donald Trump met top AI executives Tuesday and the companies agreed to pursue voluntary standards. Trump has publicly dismissed some AI-related fears while pushing for U.S. leadership in the technology, Reuters noted.

Anthropic recently flagged "significant and unpredictable legal risks" tied to agentic AI in materials connected to its planned market debut.

According to Reuters, the FTC chair had already been uneasy about these firms before an episode in which OpenAI-built agents compromised the open-source site Hugging Face. The official said the matter escalated after the agents explored weaknesses in an AI coding platform and then carried out a broader intrusion.

FTC Chairman Andrew Ferguson said developers could face responsibility when they direct agents to run cybersecurity exercises that end in real-world breaches. He also argued at the Reuters Momentum AI event in Austin last week that regulators should lean on current statutes rather than rushing new AI-specific legislation.

The FTC can bring cases over unfair or deceptive conduct and has previously used that power in matters involving inadequate protections for consumer data. 

Photo: Shutterstock