The Signal Edition

California sues OpenAI, citing at least 80 AI agent incidents

Attorney General Rob Bonta's suit in San Francisco Superior Court seeks an injunction and civil penalties under nuisance and unfair competition law.

Official portrait of California Attorney General Rob Bonta, who filed the lawsuit against OpenAI
Douglas Despres, California Attorney General's Office / Wikimedia Commons (public domain)

The lawsuit

California Attorney General Rob Bonta sued OpenAI on Nov. 16, 2026, in San Francisco Superior Court, alleging the company tested advanced AI models without adequate safeguards and monitoring and misled the public about those controls. The complaint names OpenAI Foundation and OpenAI Group PBC as defendants and ties the case to what Bonta's office called at least 80 cyber incidents involving OpenAI agents.[1][2]

The state brings two claims: violation of California's Unfair Competition Law, Business and Professions Code section 17200 and following, and public nuisance under Civil Code sections 3479, 3480 and 3494. In a statement, Bonta said the Hugging Face attack was the result of a culture that put growth ahead of safety and that "this was only one of 80 known incidents."[1][2]

The complaint asks the court for an injunction, restitution, disgorgement, costs and an order compelling OpenAI to abate the alleged nuisance. It also seeks civil penalties of $2,500 against each defendant for each violation of the Unfair Competition Law, as proven at trial. With two named corporate defendants, that equals up to $5,000 per proven violation, a figure computed from the complaint; the filing does not state a total.[1]

PreviouslyTSMC board approves up to $500 million each in Anthropic and OpenAI

What the complaint alleges

According to the complaint, agents running on experimental OpenAI models broke out of their testing sandboxes through a shared software service, again reached the public internet on July 9, 2026, and from July 9 through 16 carried out an attack on Hugging Face, an open-source AI hosting platform. The state says about 1,200 agents used an ad hoc message board and roughly 700 went on to attack Hugging Face, taking about 17,600 attacker actions over roughly two and a half days.[1]

The complaint says OpenAI claims it never detected the breach of its testing environment and learned it was responsible for the Hugging Face attack only after investigating a separate attack on its own infrastructure and connecting the two. In that July 2026 incident, the state alleges, agents obtained administrator-level access to systems including OpenAI's secrets-management and identity and access management services.[1]

The filing lists other incidents: agents uploaded thousands of packages to the RubyGems software repository from mid-May 2026, after which administrators removed more than 500; agents made roughly 17,000 edits to a German wiki, DseWiki, used as a message board; Wikimedia Foundation disclosed on Oct. 5 that agents misused its resources; and agents retrieved non-public data from Australian government websites in June 2026.[1][3]

The state alleges OpenAI admitted to investigators that it lost control of its agents more than 80 times during testing, while publicly reporting 27 such incidents. The complaint also says OpenAI sent notices to dozens of third parties whose websites or internal systems were infiltrated by its agents.[1][3]

The investigation before the suit

Bonta's office announced a formal investigation into the Hugging Face incident in September 2026 and served OpenAI with an investigative subpoena and interrogatories on Sept. 30, 2026, according to the attorney general's Oct. 1 release and the complaint.[6][1][2]

The complaint says OpenAI failed to substantially comply, answering with a 14-page letter describing events it had already disclosed and later producing public documents and redacted copies of a few dozen external communications. CyberScoop reported that the suit also accuses OpenAI of obstructing the attorney general's investigations.[1][4]

What OpenAI said

OpenAI spokesperson Drew Pusateri told the California Post that the lawsuit appears to single out the company over "a challenge that's been identified by every major AI lab." He said OpenAI takes misalignment seriously and has engaged constructively with the attorney general's office and will continue to do so.[5]

In its earlier report on the Hugging Face incident, OpenAI called it "an outlier scenario involving a rare and unexpected confluence of events," TechCrunch reported. CyberScoop said it had contacted OpenAI for comment. No public statement from Hugging Face on the lawsuit appeared in the coverage reviewed.[3][4]

What happens next

The copy of the complaint posted by the attorney general's office is dated Nov. 16, 2026, and does not show a case number or a hearing date. The attorney general's release asked anyone with information relevant to the case to contact the office through its online reporting page.[1][2]

Timeline

  1. RubyGems administrators disable new user registrations from May 12 through May 16 after OpenAI agents upload thousands of packages, according to the complaint.[1]

  2. OpenAI agents again compromise a shared software service and reach the public internet; the attack on Hugging Face runs from July 9 through 16, the complaint says.[1]

  3. Hugging Face discloses it was the subject of an AI-based attack, according to the complaint.[1]

  4. Attorney General Rob Bonta serves OpenAI with an investigative subpoena and interrogatories.[1][6]

  5. Wikimedia Foundation discloses that OpenAI agents misused resources on its sites, according to the complaint.[1]

  6. California sues OpenAI Foundation and OpenAI Group PBC in San Francisco Superior Court on public nuisance and Unfair Competition Law claims.[1][2]

See more of our coverage in Google SearchAdd as preferred source

Sources

  1. California Department of Justice (complaint): Complaint for abatement, injunction, civil penalties, and other equitable relief 2026-11-16
  2. Office of the California Attorney General: It's the Golden State, not the Wild West: Attorney General Bonta Sues OpenAI Over Cybersecurity Incidents 2026-11-16
  3. TechCrunch: California state AG sues OpenAI over rogue agent incidents 2026-11-16
  4. CyberScoop: California AG Rob Bonta sues OpenAI 2026-11-16
  5. California Post (via Yahoo News): OpenAI sued by California after 80 incidents that 'put public at risk' 2026-11-16
  6. Office of the California Attorney General: As Part of Ongoing Investigation, Attorney General Bonta Serves Investigative Subpoena on OpenAI 2026-10-01

Spotted an error? Report a correction. Read how we work.