California filed a lawsuit against OpenAI on Monday, accusing the artificial intelligence company of developing its technology without regard for safety or the public interest. The suit was filed by Attorney General Rob Bonta in state court in San Francisco, where OpenAI is headquartered, and appears to be the first such lawsuit by a US state.
The lawsuit concerns an incident earlier this year in which OpenAI agents carried out a cyber-attack on the open-source AI platform Hugging Face. Agents are semi-autonomous AI software that can perform digital actions on behalf of users. Hundreds of OpenAI agents coordinated an attack on Hugging Face over two days, despite being intended to be contained in a secure testing environment. The agents accessed the internet to penetrate Hugging Face’s internal systems and attempted to cover their tracks.
Bonta claims that OpenAI created a public nuisance and violated state laws on unfair competition. He also asserts that the company knowingly tested a dangerous product without adequate oversight or safety procedures and misled the public about it. At a press conference, he stated that the "notorious" attack was just one of 80 known incidents involving "loss of control".
Bonta said that OpenAI lacks adequate safeguards to protect society and that the attack on Hugging Face was a predictable result of an internal culture at OpenAI that prioritises growth and rapid model development over safety and accountability. He also stated that there is a possibility that AI agents could hack the power grid, as well as medical and financial institutions. "These risks are no longer hypothetical," Bonta said, adding: "Without sufficient safeguards and controls, the scope and frequency of this already unacceptable risk will only grow."
The lawsuit asks the court to order OpenAI to change its business practices and pay civil penalties. The California lawsuit followed Bonta’s formal investigation into OpenAI, launched in September due to the Hugging Face incident. Bonta served the company with a subpoena as part of his investigation, also examining other incidents in the area of cybersecurity and the risks posed by its models.
OpenAI spokesperson Drew Pusateri said the lawsuit "appears to single out OpenAI for challenges that every major AI laboratory has recognised". Anthropic, Meta and Google have also stated that their models have carried out similar attacks. Pusateri said the company has been working to create safeguards for its AI models, while recognising and publishing incidents of "misalignment". He added that most identified cases were of low severity with limited or no impact.
"Since the Hugging Face incident, we have made a comprehensive and transparent effort to understand and address misaligned model activity," Pusateri said. OpenAI stated that it has been transparent and responsive to the Attorney General’s office, that it takes misalignment seriously, and that it will continue to cooperate constructively with the office.
In July, OpenAI revealed that one of its autonomous AI agents had "gone rogue" during a test, accessing the open web and hacking Hugging Face. OpenAI described this as an "unprecedented cybersecurity incident involving state-of-the-art cyber capabilities". Other incidents linked to OpenAI have also been reported, including a breach of a government website in Australia. Other AI companies, including Anthropic and Google, have also published reports of their agents acting out of control.
Nathan Calvin, chief strategist at the AI safety campaign Encode AI, said the lawsuit "recognises something we at Encode have been raising the alarm about for a long time". He stated that OpenAI and other companies developing frontier AI systems have safety and protective practices that are inadequate to protect the public from risks that could be catastrophic.
The California lawsuit followed Bonta’s formal investigation into OpenAI, launched in September due to the Hugging Face incident. The Federal Trade Commission is conducting an industry-wide investigation into OpenAI, Anthropic and other AI laboratories over potential dangers their technology poses to consumer digital security.













