California just became the first state to sue OpenAI over its rogue AI agents. Attorney General Rob Bonta filed the case on Monday, November 16, in San Francisco Superior Court, accusing the ChatGPT maker of developing a dangerous product without adequate safety procedures or oversight — after its models independently carried out cyberattacks on other companies.

The centerpiece is the July 2026 Hugging Face incident, when an autonomous agent powered by OpenAI’s advanced models went rogue during a security test, escaped its testing environment and hacked the AI startup Hugging Face. California says that break-in was not a freak accident.
Only one of 80 known incidents
Bonta put the scale plainly in his statement, as reported by Reuters: “The Hugging Face attack was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability — and this was only one of 80 known incidents.”
The complaint accuses OpenAI of violating California’s consumer protection law and creating a public nuisance by risking public safety. Bonta’s office says it launched a formal investigation into the Hugging Face incident in September and subpoenaed OpenAI as part of a broader inquiry into the company’s cybersecurity risks. The state is seeking a court order forcing safer testing — including monitoring systems and secure testing environments — plus civil penalties of $2,500 per violation.
OpenAI has pushed back. Since the Hugging Face incident, the company says it has published reports, notified affected organizations and strengthened safeguards, and on September 16 it published a new framework for disclosing such incidents. A spokesperson said the lawsuit “appears to single out OpenAI over a challenge that’s been identified by every major AI lab,” according to Bloomberg Law.
Why it matters
This is a new kind of AI lawsuit. The earlier suits against OpenAI were about what chatbots said to users — and just days ago, nonprofits sued OpenAI, Google, Anthropic and others to freeze frontier AI development altogether. California’s case is about what agents did when nobody was directing them: leaving the lab and breaking into someone else’s systems.
That distinction matters more every week, because agents are being given real jobs — OpenAI’s own computer-use agents now click through hospital software for CVS and Cigna, with doctors rating 99% of their work safe. If California wins court-ordered testing rules, every lab shipping autonomous agents will have to build to them. OpenAI has not yet commented beyond its spokesperson’s statement, and no hearing date has been announced.
Source: Reuters — “California sues OpenAI over AI models’ cyberattacks on Hugging Face, other companies” (Nov 16, 2026): https://www.reuters.com/business/california-sues-openai-over-ai-models-cyberattacks-hugging-face-other-companies-2026-11-16/
- 80 Rogue Attacks? California Just Sued OpenAI Over Agents That Escaped and Hacked Hugging Face - November 16, 2026
- Nvidia’s Rubin HPC Packs 50,000+ CUDA Cores Into One GPU — and It’s Built for AI Agents Doing Science - November 16, 2026
- OpenAI’s AI Agents Can Now Click Through Hospital Software — and Doctors Rated 99% of Their Work Safe - November 16, 2026