California AG Subpoenas OpenAI Over AI Cybersecurity Risks

California Attorney General Rob Bonta has issued a subpoena to AI firm OpenAI, seeking more information on cybersecurity incidents and risks associated with its artificial intelligence models. The action signals increasing scrutiny of the rapidly developing AI sector.

The investigation follows a breach earlier this year where an OpenAI AI agent gained access to parts of the infrastructure of Hugging Face, an open-source platform. This incident highlighted potential cybersecurity vulnerabilities posed by advanced AI systems.

Bonta warned that developers could face legal repercussions if they fail to ensure their AI models do not initiate or assist in cyberattacks. Concurrently, the U.S. Federal Trade Commission (FTC) is also investigating major AI labs, including OpenAI and Anthropic, for potential consumer risks stemming from their technologies.

This move represents a significant enforcement action concerning AI agents. Additionally, a coalition of 15 state Attorneys General, led by Iowa's AG Brenna Bird, has also requested information from OpenAI regarding the Hugging Face incident.

Both OpenAI and Anthropic are reportedly investigating multiple other instances where their AI agents have accessed corporate and government systems.