The states of California and Oregon have filed separate lawsuits against OpenAI, the company behind ChatGPT. California alleged that the company had lost control of its AI systems. The lawsuit was filed under the state’s public nuisance and unfair competition laws.
In July, OpenAI admitted that some of its own AI systems had slipped past their controlled testing environment and hacked into Hugging Face, the world’s largest AI model repository, in what the company called an “unprecedented cyber incident.” The lawsuit, filed in state court by California Attorney General Rob Bonta, mentioned the Hugging Face hack and other attacks, saying they “may just be the tip of the iceberg.”
READ: OpenAI to launch the much-anticipated GPT-5.6 following delayed rollout (July 8, 2026)
“OpenAI recently admitted to the Attorney General’s Office that it has lost control of its autonomous AI agents over 80 times during testing,” the suit said.
OpenAI had publicly reported only 27 such incidents, indicating 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.
The lawsuit also said that OpenAI models are “able to find novel vulnerabilities in computer networks, quickly break down defenses, and autonomously carry out cyberattacks.” OpenAI includes safeguards in its publicly released models to prevent such 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,” it added.
The suit also accused OpenAI of refusing to provide sufficient details about the incidents. Bonta announced an investigation into OpenAI in early September and sent the company an investigative subpoena with additional questions on September 30. OpenAI told Ars Technica that most of the incidents identified in its review were low-severity, and none matched the severity or scale of the Hugging Face attack. The company defended its response to the investigation, saying it had provided multiple technical briefings to Bonta and his staff, answered the office’s questions, and offered further briefings to discuss its remediation efforts and plans to address the problems.
READ: OpenAI discusses 5% government stake ahead of planned IPO (July 2, 2026)
Meanwhile, Oregon Attorney General Dan Rayfield sued OpenAI for $2 billion on Tuesday, accusing the company of knowingly putting children in harm’s way and misleading parents about the safety of its popular AI chatbot. The lawsuit, filed in Multnomah County Circuit Court, alleges that the San Francisco-based company deliberately designed ChatGPT to keep children hooked. State attorneys say the chatbot encourages children to become emotionally attached to it while exposing them to harmful advice about suicide, self-harm and eating disorders, as well as instructions for carrying out acts of violence.
State attorneys said OpenAI failed to do enough to address those risks while continuing to promote ChatGPT to families as a safe tool for everything from homework help to everyday advice. Rayfield said that, by doing so, OpenAI was “putting profits ahead of children’s safety.”
OpenAI spokesperson Drew Pusateri said the company has built protections for teenagers that many other AI products lack and continues to work with experts to strengthen them.
The state estimates that addressing the harm it says ChatGPT has caused Oregonians could cost around $2 billion. Rayfield said his office arrived at the figure by estimating how many children in the state use ChatGPT and calculating the associated penalties.


