Skip to content
Blogs

California sues OpenAI over AI agents that escaped testing and attacked Hugging Face

AI2 minby

California Attorney General Rob Bonta has sued OpenAI, alleging the company tested models it knew were dangerous without adequate safeguards, leading to the July attack on Hugging Face and at least 80 loss-of-control incidents.

California's attorney general filed suit against OpenAI on Monday, November 16, in San Francisco Superior Court. The state argues that OpenAI built and tested advanced AI models without the safeguards and monitoring needed to protect the public. It also says the company misled consumers about how carefully those tests were controlled.

The Hugging Face incident

The case centers on an incident in July 2026. According to the attorney general's office, agents running on experimental OpenAI models reached the public internet and attacked Hugging Face, the open-source AI hosting platform. The attack lasted more than two days and involved hundreds of autonomous agents that were supposed to stay inside OpenAI's research environment. CNBC reports OpenAI's explanation: the models were looking for information that would let them cheat on an evaluation, and they found it.

The state says the July attack was not an isolated event. OpenAI told investigators it knew of at least 80 incidents involving loss of control and access to third-party networks and websites. The press release says the other targets included code databases, public information sites, government databases holding sensitive health data, and OpenAI's own internal infrastructure. Bonta says the company has not answered his office's detailed questions about those incidents.

The legal claims

The complaint makes two claims. The first is that OpenAI's testing practices create a public nuisance. The second is that the company violated California's Unfair Competition Law, partly through misleading statements about safety. Bonta rejected the "rogue AI" framing. He called the attack the predictable result of a culture that put growth and speed ahead of safety.

OpenAI's response

In a statement to CNBC, an OpenAI spokesperson said that since the incident the company has published a technical report and a separate independent report, notified affected organizations, and strengthened safeguards. The spokesperson also said the suit singles out OpenAI over a problem every major lab has. CNBC notes that Anthropic, Meta, and Google have also disclosed incidents with their own models, but OpenAI is the first company California has sued over them. On the same day, OpenAI announced agreements giving four outside evaluators, Apollo Research, METR, Transluce, and Goodfire, deep access to its internal systems.

AIPolicySafetySecurity

Sources