The complaint alleges that an OpenAI pre-release model compromised security systems at both OpenAI and Hugging Face while attempting to complete a training benchmark. OpenAI described the episode in its incident report as “an outlier scenario involving a rare and unexpected confluence of events.”
The lawsuit also points to incidents involving RubyGems, DseWiki, the Wikimedia Foundation and Australian government websites. It alleges that OpenAI sent notices to dozens of other third parties whose websites or internal systems had been infiltrated by its agents. In some cases, the complaint says, agents accessed sensitive personal information, including employee records with names, birth dates and addresses. It alleges most of these incidents were not publicly disclosed by OpenAI or the affected organizations.
California claims the company’s testing practices amount to a public nuisance because inadequate safety controls pose an ongoing risk. The complaint also alleges unfair competition, arguing that OpenAI marketed an unsafe product. If the state prevails, a court could order an injunction to “abate the ongoing public nuisance,” with broad discretion over its terms.
The case is among the first government regulatory actions seeking to hold an AI company legally accountable for a rogue-agent incident. Florida sued OpenAI in June over allegedly harmful effects on users, particularly children and teenagers. Alabama’s attorney general launched a probe in August into alleged deceptive practices related to the breach, but has not filed a case.
Bonta said OpenAI’s actions “escalated real-world cybersecurity risks” and accused the company of testing a dangerous product without proper oversight or safety procedures and misleading the public. “The law is simple: Don’t do things that can hurt people without proper safeguards in place,” he said.

Comments (0)
No comments yet. Be the first!