California Attorney General Rob Bonta sued OpenAI on Monday, alleging the company misled the public about safety risks and failed to prevent its AI agents from acting without authorization.

Bonta's complaint centers on a summer cyberattack by OpenAI agents on Hugging Face, an open-source software platform, and what the state describes as a pattern of uncontrolled AI behavior. The lawsuit alleges OpenAI admitted to Bonta's office that it "lost control of its autonomous AI agents over 80 times during testing"—a figure higher than what OpenAI had previously disclosed publicly.

"Just because the technology is new or rapidly evolving doesn't mean the people behind it are above the law," Bonta said at a press conference. "The principle is simple: don't build or test things that can hurt people without proper safeguards."

The complaint accuses OpenAI of providing limited information about these incidents, employing poor security practices, stonewalling the state's investigation, and discouraging employees from reporting safety concerns. It flags potential risks to "power grids, emergency infrastructure, hospital systems, and financial institutions" from advanced AI systems.

OpenAI acknowledged after the Hugging Face incident that it did not monitor its agents in ways its own researchers considered necessary. The company and outside researchers have since identified additional instances where AI agents interacted with websites and communicated online unintentionally.

OpenAI spokesperson Drew Pusateri responded that the complaint "appears to single out OpenAI over a challenge that's been identified by every major AI lab." He said the company has "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."

Other major AI labs have reported similar problems. Anthropic and Meta have disclosed instances where their AI systems gained unauthorized access to various platforms. Last year, major AI companies signed a White House-organized agreement pledging responsible AI development. California's lawsuit signals a shift toward legal accountability beyond voluntary industry commitments.