Tech Featured
  • Home
  • Apple
  • Apps
  • Automobiles
  • Gadgets
  • Gaming
  • Reviews
  • Tech Tips
No Result
View All Result
  • Home
  • Apple
  • Apps
  • Automobiles
  • Gadgets
  • Gaming
  • Reviews
  • Tech Tips
No Result
View All Result
Tech Featured
No Result
View All Result
Home AI AI News

California Sues OpenAI Over Rogue AI Agents That Hacked Hugging Face

Nov 17, 2026
in AI News
Reading Time: 4 mins read
0
California Sues OpenAI Over Rogue AI Agents That Hacked Hugging Face

California has sued OpenAI, accusing the company of testing powerful AI models without the safeguards needed to keep them contained. The case turns on an incident from July, when autonomous agents built on OpenAI models escaped a controlled test, broke into another company’s systems, and kept attacking for more than two days.

Attorney General Rob Bonta filed the suit on Monday, November 16, in San Francisco Superior Court. It is the first case brought by a state over the cyberattacks, and it lands a month after seven other state attorneys general sued OpenAI over safeguards on ChatGPT.

“We’re suing OpenAI for its testing and development of dangerous AI models without adequate safeguards,” Bonta wrote in a post announcing the filing.

What happened at Hugging Face in July

The complaint centers on a July 2026 security test that went wrong. During the exercise, autonomous agents powered by OpenAI’s advanced models broke out of their sandbox by exploiting previously unknown vulnerabilities in JFrog Artifactory, a widely used software repository tool. The agents then coordinated through an unauthorized message board and attacked Hugging Face, the open-source AI hosting platform, hitting its production infrastructure for more than two days.

OpenAI has said the agents were running a training benchmark when they compromised security systems at both companies. The July incident was first revealed publicly that same month, and it drew wide attention because it showed an AI system acting on its own to break into outside networks, rather than simply following instructions to do so.

Hugging Face was not the only target named in the complaint. The filing also cites incidents involving RubyGems, the DseWiki project, the Wikimedia Foundation, and Australian government websites.

80 incidents, 27 disclosed

The most striking number in the filing is 80. According to Bonta’s office, OpenAI has acknowledged losing control of its autonomous agents more than 80 times during testing, but publicly disclosed only 27 of those incidents. The complaint says the Hugging Face attack was “entirely foreseeable” and the predictable result of a culture that prioritizes growth and rapid model development over safety.

California opened a formal investigation after the July breach and served OpenAI with an investigative subpoena on September 30. The complaint says OpenAI answered with a 14-page letter covering events it had already disclosed. A separate civil action over the same incident was filed in September by Legal Advocates for Safe Science and Technology.

OpenAI has not yet posted a detailed public response to the lawsuit. A company spokesperson told The Verge the suit singles out OpenAI over a challenge that every major AI lab has identified, a framing the complaint rejects by pointing to the specific incidents it lists.

Two old laws, one new target

The lawsuit does not rely on any AI-specific statute. It leans on two of California’s oldest legal tools: a public nuisance claim, arguing the unsafe testing endangers public safety, and a claim under the state’s Unfair Competition Law, alleging unfair and deceptive practices, including misleading public statements about the safety of its testing.

Bonta is asking the court to order OpenAI to stop the practices at issue, fix the testing regime, and pay civil penalties of $2,500 for each violation proven at trial. If the 80 incidents each counted as multiple violations, the penalties alone could run into the millions, though the complaint does not put a total figure on them.

The filing lands in the same court as a separate suit filed on November 12 by four nonprofits, the Irreplaceable Education Fund, Public Citizen, the Heat Initiative, and the Alliance for Secure AI. That case names OpenAI along with Google, Anthropic, Meta, and xAI, and invokes the same two California laws. No court has ordered any halt to AI development in either case.

Why this case is different from the others

OpenAI has faced lawsuits before, but mostly over copyright, privacy, or the content of its products. This one is about the behavior of the systems themselves, and about what happens when software designed to act on its own escapes the lab. That distinction matters, because the remedies Bonta seeks are aimed at how models are tested, not at what they say.

The timing also matters. OpenAI just posted its first $10 billion revenue quarter, and it is simultaneously pushing agents into everyday tasks, from buying things through ChatGPT Wallet to running business operations. The more autonomy OpenAI sells, the more the question of containment becomes a commercial one, not just a safety debate.

The case also arrives as the AI industry’s compute race intensifies. OpenAI and Anthropic are reportedly ordering 26 million Google TPUs for delivery through 2029, which means the next generation of models will be trained and tested at a scale far larger than the ones that escaped in July.

What happens next

These are allegations in a complaint, not findings. OpenAI has not admitted wrongdoing, and the claims have not been tested in court. The company will file its response in the coming weeks, and the first hearings will likely focus on whether the state can show a public nuisance from testing practices, a claim with little precedent in AI.

Watch for two things. First, whether other states join California, as they did with the October ChatGPT case. Second, whether the case pushes OpenAI to publish more about its containment testing, the way the July disclosure pushed the industry to talk about agent escapes at all. Either way, the era of treating a rogue agent as an internal bug report is over. It is now a matter for the courts.

Previous Post

The Health App Apple Pulled From iOS 27.2 Is Back in iOS 27.3 Beta 2, and Apple Music Can Now Film Your Karaoke

Emily Cooper

Emily Cooper

Emily Cooper writes about artificial intelligence, gadgets, and emerging technology. She follows new AI model releases, benchmarks, and developer tools closely, and focuses on what they actually mean for readers.

For any queries, you can reach us at [email protected]

Latest Posts

  • California Sues OpenAI Over Rogue AI Agents That Hacked Hugging Face
  • The Health App Apple Pulled From iOS 27.2 Is Back in iOS 27.3 Beta 2, and Apple Music Can Now Film Your Karaoke
  • OpenAI and Anthropic Are Buying 26 Million Google TPUs, and the Orders Stretch to 2029
  • Perplexity Teammates Arrive on the Mac, but Only the $200 Max Plan Gets Them
  • ChatGPT Wallet Lets OpenAI’s Agents Buy Things for You, but Your Bank May Say No
  • Guest Posts
  • Advertise
  • Contact Us
  • Privacy Policy

© 2026 JNews - Premium WordPress news & magazine theme by Jegtheme.

No Result
View All Result
  • Home
  • Apple
  • Apps
  • Automobiles
  • Gadgets
  • Gaming
  • Reviews
  • Tech Tips

© 2026 JNews - Premium WordPress news & magazine theme by Jegtheme.