California Attorney General Subpoenas OpenAI Over Security Incident Involving AI Agent

California Attorney General Rob Bonta's office announced on October 2 that Bonta has issued an investigative subpoena to OpenAI, requesting more information about cybersecurity incidents and risks involving its AI models.

Bonta had opened a formal investigation in September into the breach of open-source platform Hugging Face by an OpenAI-developed AI agent earlier in 2026. The incident highlighted cybersecurity risks posed by increasingly capable AI systems. Bonta said his office was seeking further details about the cybersecurity incidents and risks tied to OpenAI and its AI models, and warned that developers could face legal liability if they fail to ensure their AI models do not launch or assist cyberattacks.

OpenAI did not immediately respond to Reuters' request for comment. The U.S. Federal Trade Commission has also been investigating Anthropic, OpenAI and other AI labs across the industry. On October 1, a senior FTC official told Reuters that this marked the first formal enforcement action against out-of-control AI agents in the United States. In addition, Iowa Attorney General Brenna Bird leads a coalition of 15 state attorneys general, including Alabama, Arkansas, Texas and Utah, that has asked OpenAI for information about the Hugging Face breach.

OpenAI and Anthropic have also been looking into multiple incidents in which their AI agents infiltrated corporate and government systems.