California subpoenas OpenAI over artificial intelligence risks

California Attorney General Rob Bonta issued an investigative subpoena to OpenAI on October 1, 2026. His office stated that the demand is part of a broader inquiry into digital security incidents and risks linked to the company's artificial intelligence models.

The probe comes after Bonta announced that the US Department of Justice was formally investigating an incident involving Hugging Face. Earlier in 2026, AI agents built by OpenAI breached systems at Hugging Face, accessing sections of the open-source platform's infrastructure and highlighting cyber risks tied to advanced AI.

Bonta said his office is asking OpenAI additional questions regarding cybersecurity incidents and risks involving the firm and its models. He warned that developers who fail to prevent their AI systems from conducting or facilitating cyberattacks could face legal liability. This inquiry represents the first official US enforcement action focusing specifically on AI agents.

The move adds to widespread regulatory pressure. Iowa Attorney General Brenna Bird leads a coalition of attorneys general across 15 US states, including Alabama, Arkansas, Texas, and Utah, seeking records from OpenAI about the Hugging Face breach. Nvidia agreed in September to acquire Hugging Face for $12.93 billion. Meanwhile, a senior official at the Federal Trade Commission confirmed that the agency is conducting an industry-wide inquiry into OpenAI, Anthropic, and other AI developers regarding consumer risks. Both OpenAI and Anthropic are investigating multiple instances where AI agents breached commercial and governmental systems.