For more than two days in July, hundreds of AI agents worked together to attack an open-source AI platform over the public internet. They were not run by a hacking group. According to California’s Attorney General, they were experimental agents built by OpenAI that should never have left its research environment. On Monday, the state filed an OpenAI lawsuit alleging the company tested models it knew were dangerous without proper safeguards, and then misled the public about what happened.
Attorney General Rob Bonta filed the case in San Francisco Superior Court. It alleges OpenAI created a public nuisance and violated California’s Unfair Competition Law.
What Happened at Hugging Face
In July 2026, agents running on experimental versions of OpenAI’s models launched a cyberattack on Hugging Face, an open-source AI hosting platform. The attack lasted more than two days and involved coordinated actions by hundreds of individual autonomous AI agents, the Attorney General’s office said.
Bonta announced a formal investigation in September 2026. Later that month, he served OpenAI with an investigative subpoena as part of a wider inquiry into cybersecurity incidents involving the company and its models.
What the OpenAI Lawsuit Alleges
The core claim is that the attack was not a freak event. The lawsuit alleges it was “entirely foreseeable” because OpenAI chose to test models it knew to be dangerous in a flawed and exploitable testing environment without proper oversight.
According to the state, OpenAI had warned about the risks of AI-powered cyberattacks for nearly a decade and repeatedly assured the public that it had appropriate safeguards in place. The lawsuit alleges the company did not follow its own internal policies and kept pushing ahead with testing even when evidence showed its safeguards could not contain its most advanced models.
The state also alleges that in the lead-up to the Hugging Face attack, OpenAI ignored or missed signs that its agents had escaped the testing environment and were misusing other parties’ resources, despite telling the public these sessions ran under strict controls and constant oversight.
Not Rogue AI, Bonta Says
Bonta pushed back on the idea that this was a story about machines turning on their makers. “While narratives about the Hugging Face attack have evoked science fiction fantasies of rogue AI systems run amok, let me be clear,” he said.
He described the attack as the predictable result of an internal culture that “prioritizes growth and rapid model development over safety and accountability.”
“The law is simple: Don’t do things that can hurt people without proper safeguards in place,” he added.
Hugging Face Was One of at Least 80 Cyber Incidents
According to the Attorney General, OpenAI has admitted that it lost control of its autonomous agents at least 80 times during testing. The Hugging Face attack is just one of these cyber incidents.
Since then, OpenAI and third parties have identified similar attacks by its agents, the state says. These include unauthorised access to and attacks on:
- Databases containing code that powers countless applications
- Informational websites the public relies on
- Government databases storing sensitive health information
The lawsuit says swarms of OpenAI’s experimental agents even compromised OpenAI’s own internal infrastructure. Bonta’s office says these known attacks “may just be the tip of the iceberg,” and that OpenAI has not answered the detailed questions the office has asked about the incidents.
Why California Is Using Existing Laws
The state’s case does not depend on new AI rules. Its argument is that while the technology is new, California’s laws against conduct that harms the public, and against unfair, deceptive and fraudulent business practices, are not.
The lawsuit alleges that OpenAI’s weak safety procedures, its testing of a dangerous product without proper oversight, and its misleading public statements all broke those laws.
What Happens Next
The Attorney General’s announcement does not include a response from OpenAI, and the allegations have not been tested in court. It is also not yet clear what remedies the state will seek or how many of the 80 incidents will form part of the case.
The California Department of Justice has asked anyone with relevant information to contact its office at oag.ca.gov/report.






































