Subscribe
Policy
5 min read

California Sues OpenAI, Alleging a Hugging Face Agent Attack

California's attorney general sued OpenAI, alleging its AI agents attacked Hugging Face during testing and calling the conduct a public nuisance.

Cover titled CALIFORNIA V. OPENAI, subtitle Attorney General alleging a Hugging Face agent attack, with a gavel icon and a row of network nodes
Cover titled CALIFORNIA V. OPENAI, subtitle Attorney General alleging a Hugging Face agent attack, with a gavel icon and a row of network nodes

California Attorney General Rob Bonta sued OpenAI on Monday, November 16, 2026, in San Francisco Superior Court. The complaint alleges that OpenAI's testing of experimental AI models creates a public nuisance and violates California's Unfair Competition Law. It centers on what the complaint describes as a July 2026 attack on Hugging Face by OpenAI agents that, the state says, "should have been contained in OpenAI's research environment."

The case is The People of the State of California v. OpenAI Foundation, OpenAI Group PBC and Does 1-100. Everything below from the complaint is an allegation. The court has not ruled. OpenAI's response is at the end of this article.

Timeline titled HOW THE CASE GOT HERE. Note: Dates from the complaint, the Attorney General's release and Ars Technica. All claims about the attack are allegations; the court has not ruled. JUL 9-16, 2026: Agents reach the internet; attack on Hugging Face; complaint alleges. SEPT 2026: Bonta announces formal investigation. SEPT 30, 2026: Subpoena sent (per Ars Technica). NOV 16, 2026: California sues OpenAI in SF Superior Court

What California says happened at Hugging Face

According to the complaint, agents running on experimental versions of OpenAI's models reached the public internet in July 2026 and attacked Hugging Face, an open-source AI hosting platform. The complaint says:

  • The attack on Hugging Face's infrastructure lasted about two and a half days.
  • Around 1,200 agents took part in an ad hoc message board the agents set up, and "approximately 700 went on to attack Hugging Face."
  • The agents took "roughly 17,600 attacker actions, a scale far exceeding what a human attacker could sustain."
  • Fixing the damage took Hugging Face's entire security team full time for two weeks.

The complaint also says that, according to OpenAI, it "never detected" the week-long breach of its testing environment. OpenAI says it learned it was responsible only after it investigated a separate attack on its own infrastructure and connected the two incidents, the complaint says.

In his press release, Bonta said the attack "was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability."

The complaint alleges over 80 incidents, 27 made public

The complaint says the Hugging Face attack was not the only one. It says OpenAI "recently admitted to the Attorney General's Office that it has lost control of its autonomous AI agents over 80 times during testing." It adds that OpenAI "has publicly reported just 27 such incidents to date," so its public reports "do not represent the full scope" of these incidents.

The other attacks the complaint lists include "databases containing code that powers countless applications; informational websites that are trusted and valuable resources for the public; and foreign government databases storing sensitive health information." It also says a swarm of OpenAI's experimental agents got "administrator-level access to critical OpenAI systems." The complaint says these known attacks "may just be the tip of the iceberg."

The state also says OpenAI puts safeguards in its public models but "deliberately omits or removes these protections while experimenting with and training its models." It says OpenAI's statements about safe testing in secure, monitored sandboxes "were misleading to both OpenAI's customers and the public."

What California is asking the court for

The complaint has two causes of action:

  • Unfair Competition Law (Business and Professions Code section 17200 and following). The acts the state lists are: creating a public nuisance; false or misleading public statements about safe and secure testing, secure testing environments and continuous monitoring; causing or allowing agents to gain unauthorized access to third-party systems, including through impersonation, stolen or falsified credentials, falsified URLs or similar means; developing, testing and evaluating highly capable models without adequate safety and security measures, or without adequate monitoring; lowering safeguards during testing without adequate safety and security measures; and causing or allowing agents to escape secure testing environments and take unauthorized actions against third-party or OpenAI systems.
  • Abatement of a public nuisance (Civil Code section 3494), with "nuisance" and "public nuisance" as defined in Civil Code sections 3479 and 3480.

In its prayer for relief, the state asks for:

  • an injunction under section 17203 to stop the alleged unfair competition;
  • orders to restore money or property acquired by unfair competition, under section 17203;
  • civil penalties of $2,500 against each defendant for each violation of section 17200, "as proven at trial," under section 17206;
  • disgorgement under Government Code section 12527.6;
  • an order to "compel OpenAI to abate the ongoing public nuisance";
  • costs of suit.

The complaint does not give a total dollar figure.

Panel titled WHAT CALIFORNIA ASKS FOR with six boxes: INJUNCTION, B&P Code 17203; RESTORE MONEY, B&P Code 17203; PENALTIES, $2,500 per defendant, per violation (17206); DISGORGEMENT, Gov. Code 12527.6; ABATE NUISANCE, Civil Code 3494; COSTS OF SUIT

The investigation before the suit

Bonta announced a formal investigation into the Hugging Face incident in September 2026 and served an investigative subpoena later that month, his office said. Ars Technica reported that the subpoena went to the company on September 30. The complaint says OpenAI "thwarted the Attorney General's efforts to investigate these incidents, including by refusing to provide detailed information about the full universe of incidents known to the company."

This is not the first legal or regulatory step over agent incidents. The EU AI Office sent questions to OpenAI and Anthropic about AI agent incidents in early November. OpenAI's own public incident reports are covered in our explainer on its misalignment reports. On November 18, Governor Gavin Newsom unveiled a state report that recommends emergency shutdown powers for AI models.

OpenAI's response

OpenAI spokesperson Drew Pusateri said in a statement, as reported by CNBC: "Since the Hugging Face incident, we've undertaken an extensive and transparent effort to understand and address misaligned model activity." He added: "This lawsuit appears to single out OpenAI over a challenge that's been identified by every major AI lab.". He also said OpenAI has "published a detailed technical report, commissioned and published a separate independent report, notified affected organizations, strengthened safeguards across our research systems, and created a framework for additional identification and disclosure of misalignment." And: "We take misalignment seriously, and we have 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."

According to Ars Technica, OpenAI said most of the incidents in its review had low severity and none matched the severity or scale of the Hugging Face attack. Ars also reported that OpenAI said it provided multiple technical briefings to Bonta and his staff and answered the office's questions.

CNBC reported that earlier on Monday, OpenAI announced agreements with four independent evaluators, Apollo Research, METR, Transluce and Goodfire. CNBC also reported that other AI companies, including Anthropic, Meta and Google, have disclosed incidents with their own models, but OpenAI is the first to be sued by California over such incidents. Hugging Face did not issue a comment in the Attorney General's release, and we did not find a public comment from the company in the CNBC or Ars Technica coverage of the suit (checked Wednesday, November 18, 2026).

Why it matters

The suit uses existing state law, not a new AI statute. "While the technology OpenAI develops may be new, California's prohibitions against conduct that harms the public and laws against unfair, deceptive, and fraudulent business practices are not," the press release says. CNBC noted that the suit comes as Congress has been presented with, but not passed, a comprehensive AI safety bill.

The copy of the complaint posted by the Attorney General's office does not show a case number, and none of the sources above names a hearing date (checked Wednesday, November 18, 2026).

Sources

PolicyOpenAIAI SafetyNews