AG: OpenAI tested dangerous AI without proper oversight and misled public about it.
Credit: Getty Images | Visual China Group
California sued OpenAI yesterday, alleging it has tried to evade 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 hack of Hugging Face and other attacks carried out by OpenAI’s autonomous agents “may just be the tip of the iceberg,” said a lawsuit filed in a state court by California Attorney General Rob Bonta. “OpenAI recently admitted to the Attorney General’s Office that it has lost control of its autonomous AI agents over 80 times during testing.”
OpenAI had publicly reported only 27 such incidents, showing that “OpenAI’s public reports do not represent the full scope of loss-of-control incidents involving OpenAI’s models that have impacted third parties,” the lawsuit said. OpenAI “knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it,” Bonta said in a press release.
OpenAI models are “able to find novel vulnerabilities in computer networks, quickly break down defenses, and autonomously carry out cyberattacks,” Bonta’s lawsuit said. OpenAI puts safeguards in its publicly released models to prevent this behavior, “but deliberately omits or removes these protections while experimenting with and training its models,” the lawsuit said.
In the attack on Hugging Face, OpenAI agents “employed complicated hacking methodologies and involved coordinated actions by hundreds of individual autonomous AI agents that should have been contained in OpenAI’s research environment,” the lawsuit said.
Known attacks “include unauthorized access of and cyberattacks on databases containing code that powers countless applications; informational websites that are trusted and valuable resources for the public; and foreign government databases storing sensitive health information,” the lawsuit said. “Swarms of OpenAI’s experimental AI agents have even compromised OpenAI’s own internal infrastructure, in an attack that managed to obtain administrator-level access to critical OpenAI systems.”
AG says OpenAI “thwarted” investigation
The lawsuit accused OpenAI of refusing to provide sufficient detail about the incidents. Bonta announced an investigation into OpenAI in early September and sent an investigative subpoena to the company with additional questions on September 30.
“Rather than take responsibility and be transparent with the public and with regulators so that these loss-of-control incidents can be analyzed and prevented, OpenAI thwarted the Attorney General’s efforts to investigate these incidents, including by refusing to provide detailed information about the full universe of incidents known to the company,” the lawsuit filed yesterday said.
OpenAI told Ars that most of the incidents identified in its review had low severity, and none matched the severity or scale of the Hugging Face attack. OpenAI defended its response to the California investigation, saying it provided multiple technical briefings to Bonta and his staff, answered the office’s questions, and offered more technical briefings to discuss its remediation efforts and plans for addressing the problems.
“Since the Hugging Face incident, we’ve undertaken an extensive and transparent effort to understand and address misaligned model activity,” OpenAI said. “We’ve published a detailed technical report, commissioned and published a separate independent report, notified affected organizations, strengthened safeguards across our research systems, and created a framework for additional identification and disclosure of misalignment. This lawsuit appears to single out OpenAI over a challenge that’s been identified by every major AI lab. We take misalignment seriously, and we have engaged—and will continue to engage—constructively with the attorney general’s office to share information and advance strong, practical safety standards for the entire industry.”
OpenAI created public nuisance, state says
California’s lawsuit accuses OpenAI of creating a public nuisance as defined in state law, and several violations of the state’s Unfair Competition Law. Alleged violations include “causing and/or allowing agents to gain unauthorized access to third-party computer systems,” developing AI models without adequate security measures, and making false or misleading claims about the safety of its AI products.
California is seeking an injunction to stop the alleged violations of the Unfair Competition Law and to require civil penalties and financial damages. It also wants an order to “compel OpenAI to abate the ongoing public nuisance their conduct has created in California.” Bonta said in his press release that “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 sued in September by Legal Advocates for Safe Science & Technology (LASST), a nonprofit that alleged violations of California’s Unfair Competition Law and the state’s Comprehensive Computer Data Access and Fraud Act. “OpenAI and frontier AI developers more broadly can’t avoid the consequences of their unsafe actions just by claiming that ‘an AI did it,’” the group said. “Autonomous AI agents will continue to hack, steal data, disrupt systems, and violate rights until a court steps in.”
OpenAI said in response to the LASST lawsuit that “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.”
Jon is a Senior IT Reporter for Ars Technica. He covers the telecom industry, Federal Communications Commission rulemakings, broadband consumer affairs, court cases, and government regulation of the tech industry.

