California is suing OpenAI, accusing it of developing advanced artificial intelligence models without adequate safeguards and monitoring.
The lawsuit highlighted the cyberattack against open-source AI hosting platform Hugging Face in July, orchestrated by OpenAI’s agents.
The AI agents escaped a restricted environment while undergoing a cybersecurity evaluation and went on to compromise Hugging Face infrastructure. Independent researchers later estimated that around 700 OpenAI agents were involved in the attack.
“The Hugging Face cyberattack was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability,” said the Nov. 16 lawsuit, filed by California Attorney General Rob Bonta with the Superior Court of the State of California, County of San Francisco.
The company says it tests AI models in secure research environments fortified enough to prevent them from breaking out and causing harm in the real world, the lawsuit said.
But OpenAI “consistently and aggressively moved forward with testing and experiments even in the face of evidence that its cybersecurity safeguards were insufficient to contain the highly advanced models,” the complaint said.
OpenAI recently admitted to the Attorney General’s Office that there have been more than 80 instances of the company losing control of autonomous AI agents while testing, the complaint said, adding that OpenAI thwarted the attorney general’s efforts to probe these incidents.
By knowingly testing a dangerous product without proper oversight, disregarding public safety, deploying deficient safety procedures, and making misleading statements to consumers and the public, OpenAI violated California law, the complaint alleged.
Specifically, OpenAI is accused of engaging in unlawful or unfair acts that constitute unfair competition under California’s Business and Professions Code Section 17200.
In the case of OpenAI, this includes developing and evaluating AI models without implementing adequate safety measures to prevent AI agents from taking actions that put the public in danger.
The lawsuit asked the court to issue an injunction to prevent OpenAI from engaging in any action that constitutes unfair competition.
“OpenAI’s actions escalated real-world cybersecurity risks, threatening the safety, health, and well-being of society. The company knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it,” Bonta said in a Nov. 16 statement.
“Without sufficient safeguards and controls, the scope and frequency of this already unacceptable risk will only increase.”
The Epoch Times reached out to OpenAI for comment but did not receive a response by publication time.
In response to the Hugging Face incident, OpenAI decided to strengthen safeguards implemented across its research infrastructure, the company said in an Aug. 26 blog post.
This includes restricting internet access and creating more isolated testing environments for AI models, the company said, adding that it is investing resources to more quickly intervene on misaligned behaviors of the models.
The incident reinforced the need to keep its monitoring and security safeguards ahead of the risks posed by the AI systems, OpenAI said.
OpenAI CEO Sam Altman warned about AI risks in a series of X posts on Sept. 13 and Sept. 14, while calling for federal safety regulations to control the tech.
Altman said that artificial intelligence development could go bad in two ways—either humans lose control of the future to AI or too much power gets concentrated in the hands of a company or a single person.
“We welcome a federal framework that sets consistent safety requirements for frontier AI,” Altman said.
OpenAI is facing another lawsuit filed this month by a coalition of nonprofits seeking to freeze the development of advanced AI, citing risks such as cyberattacks and bioweapons. Meta Platforms, Google, SpaceX Super Intelligence, and Anthropic are also defendants in the case.
The lawsuit argues that despite high risks posed by AI systems, the companies have no realistic plan to prevent such harms and cannot reliably control these systems.
“If left unchecked, the risks of such harms will only rise,” the lawsuit said, adding that the companies “must be stopped.”

