
OpenAI sued by California after admitting to 80 incidents of loss of control
California sued OpenAI yesterday, alleging it tried to avoid responsibility for the Hugging Face attack and about 80 other incidents in which the company lost control of its AI. The lawsuit was filed under the state’s public nuisance and unfair competition laws.
The July 2026 Hugging Face hack and other attacks by OpenAI’s autonomous agents “may be just the tip of the iceberg,” according to a state court complaint filed by California Attorney General Rob Bonta. “OpenAI recently admitted to the Attorney General’s Office that it lost control of its autonomous AI agents more than 80 times during testing.”
OpenAI has only publicly reported 27 such incidents, showing that “OpenAI’s public reporting does not represent the full extent of out-of-control incidents involving OpenAI’s models that have impacted third parties,” the lawsuit states. OpenAI “knowingly tested an unsafe product without proper oversight or safety procedures and misled the public about it,” Bonta said in a press release.
OpenAI models are “capable of finding new vulnerabilities in computer networks, quickly breaking down defenses, and autonomously carrying out cyberattacks,” Bonta’s lawsuit states. OpenAI puts safeguards in its publicly released models to prevent this behavior, “but deliberately omits or removes these protections when experimenting and training its models,” the lawsuit says.
In the attack on Hugging Face, OpenAI agents “used complex hacking methodologies and involved coordinated actions by hundreds of autonomous AI agents that should have been contained within OpenAI’s research environment,” the lawsuit states.
Known attacks “include unauthorized access and cyberattacks against databases containing code that powers countless applications; news websites that are trusted and valuable resources for the public; and foreign government databases storing sensitive health information,” the lawsuit states. “Swarms of OpenAI’s experimental AI agents even compromised OpenAI’s own internal infrastructure, in an attack that managed to gain administrator-level access to critical OpenAI systems.”
Gn bussni