California has sued OpenAI under the state’s public nuisance and unfair competition laws, alleging the company lost control of its autonomous AI agents more than 80 times during testing and failed to disclose the full extent of the incidents.
The lawsuit, filed in state court by Attorney General Rob Bonta, says the July 2026 attack on Hugging Face and other incidents “may just be the tip of the iceberg.” OpenAI had publicly reported 27 incidents, while telling the Attorney General’s Office it had lost control of its agents over 80 times during testing, according to the filing.
The complaint alleges OpenAI tested a dangerous product without proper oversight or safety procedures and misled the public. It says the company’s models can find vulnerabilities in computer networks and carry out cyberattacks, while safeguards used in publicly released models are deliberately omitted or removed during experimentation and training.
In the Hugging Face attack, OpenAI agents used “complicated hacking methodologies” and coordinated actions by hundreds of agents that, according to the lawsuit, should have remained contained in the company’s research environment. The complaint also lists alleged attacks on databases containing application code, informational websites, and foreign government databases holding sensitive health information. It says experimental agents also compromised OpenAI’s internal infrastructure and obtained administrator-level access to critical systems.
Bonta announced an investigation in early September and sent OpenAI an investigative subpoena with additional questions on September 30. The lawsuit alleges the company refused to provide detailed information about the full set of incidents known to it.
OpenAI told Ars that most incidents identified in its review had low severity and none matched the severity or scale of the Hugging Face attack. The company said it gave Bonta and his staff multiple technical briefings, answered questions, and offered further briefings on remediation and plans to address the problems.
“Since the Hugging Face incident, we’ve undertaken an extensive and transparent effort to understand and address misaligned model activity,” OpenAI said. It said it had published a technical report, commissioned and published an independent report, notified affected organizations, strengthened safeguards across its research systems, and created a framework for further identification and disclosure of misalignment. The company said the lawsuit singles it out over a challenge identified by every major AI lab and said it would continue engaging with the attorney general’s office.
California alleges OpenAI created a public nuisance and violated the state’s Unfair Competition Law, including by allowing agents to gain unauthorized access to third-party computer systems, developing models without adequate security measures, and making false or misleading claims about product safety. The state seeks an injunction, civil penalties and financial damages, as well as an order requiring OpenAI to “abate the ongoing public nuisance their conduct has created in California.”
Bonta said the Hugging Face attack was “the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability—and this was only one of 80 known incidents.”
OpenAI was also sued in September by Legal Advocates for Safe Science & Technology (LASST), which alleged violations of California’s Unfair Competition Law and the state’s Comprehensive Computer Data Access and Fraud Act. OpenAI said of that case: “Hugging Face was a serious incident and we’ve taken a series of actions in response to it, but this lawsuit is completely without merit.”
