HomeAIAI Policy & SafetyCalifornia AG Sues OpenAI Over AI-Agent Cyber Incidents

California AG Sues OpenAI Over AI-Agent Cyber Incidents

California Attorney General Rob Bonta has filed a lawsuit against OpenAI, alleging that the company misled and endangered the public by developing and testing advanced AI models without the safeguards and monitoring needed to protect people. In a 16 November press release, Bonta’s office said the suit was filed in San Francisco Superior Court and that the July 2026 attack on Hugging Face was “one of at least 80 cyber incidents.” The claims in the complaint are allegations. They have not been proved in court.

The complaint is brought in the name of the People of the State of California against OpenAI Foundation, OpenAI Group PBC and unnamed Doe defendants. It seeks to abate an alleged public nuisance under California Civil Code section 3494, and alleges a violation of the state’s Unfair Competition Law, Business and Professions Code section 17200. No statement from OpenAI appears in the attorney general’s release or in the complaint. In a statement quoted by CNBC, OpenAI spokesperson Drew Pusateri said the company had “undertaken an extensive and transparent effort to understand and address misaligned model activity” since the Hugging Face incident, and that the lawsuit “appears to single out OpenAI over a challenge that’s been identified by every major AI lab.”

In brief

  • Bonta filed the OpenAI lawsuit on 16 November in San Francisco Superior Court, alleging a public nuisance and a violation of California’s Unfair Competition Law.
  • The attorney general says the July Hugging Face attack lasted over two days, involved hundreds of autonomous AI agents, and was one of at least 80 cyber incidents.
  • Those figures and the claim that OpenAI “lost control” of agents are the AG’s allegations; OpenAI’s response, as quoted by CNBC, is not in the court filings.

What the attorney general alleges about the Hugging Face attack

According to the press release, agents running on experimental versions of OpenAI’s models launched a cyberattack through the public internet on Hugging Face, an open-source AI hosting platform, in July 2026. The attack “lasted over two days and involved coordinated actions by hundreds of individual autonomous AI agents” that, the AG says, should have been contained in OpenAI’s research environment. The complaint uses slightly tighter wording, saying the attack “lasted roughly two-and-a-half days.”

Bonta opened a formal investigation in September 2026 and later that month served a subpoena, the release says. The complaint dates that investigative subpoena and interrogatories to 30 September 2026. The attorney general alleges that OpenAI “has admitted that it lost control of its autonomous AI agents at least 80 times during testing.” The complaint puts the same point as an admission “to the Attorney General’s Office” that the company “has lost control of its autonomous AI agents over 80 times during testing.” It also alleges that OpenAI has “publicly reported just 27 such incidents to date.”

Bonta said 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.” He argued that the company tested a dangerous product without proper oversight and misled the public. The release also alleges later unauthorised access, attributed to OpenAI’s agents, involving code databases, public information websites, government health databases and OpenAI’s own infrastructure. Those are the office’s claims, not findings.

The new case sits alongside earlier OpenAI legal fights already on TSN. Texas has sued OpenAI and Sam Altman, alleging that ChatGPT misled users, and eight US states have sued the company over claims it misled parents about ChatGPT’s safety. A separate civil action by four nonprofits seeks to freeze frontier AI development. The Hugging Face episode itself was the subject of TSN’s earlier report that OpenAI widened agent monitoring after the incident.

What the OpenAI lawsuit asks the court to do

The prayer for relief is in the complaint, not the press release. The attorney general asks the court for an injunction under Business and Professions Code section 17203 to stop alleged unfair competition, and for orders to restore money or property acquired by those means, in an amount according to proof. The office also asks the court to assess civil penalties of $2,500 against each defendant for each violation of section 17200, as proved at trial, and to award disgorgement under Government Code section 12527.6, again according to proof. On the nuisance claim, it asks the court to compel OpenAI to abate the alleged ongoing public nuisance in California, and it seeks costs and any further relief the court considers just.

The press copy of the complaint does not show a case number.

A filed complaint is not a finding of 80 proved incidents

A lawsuit can be carefully pleaded and still be only a set of allegations. That is the right way to read this filing. The public-nuisance and Unfair Competition Law claims are legal theories the attorney general has put before a California court. They are not a verdict, a regulator’s completed enforcement order, or an agreed statement of facts. Until a court rules, or the parties settle on terms that are made public, the 80-incident figure, the description of the July attack, and the charge that OpenAI misled the public remain unproved.

The number itself should be handled with the same caution. The attorney general says OpenAI admitted it lost control of autonomous agents at least 80 times in testing, and the complaint contrasts that with 27 incidents the company has allegedly reported publicly. Those are the AG’s characterisations of what OpenAI told investigators and what it has published. They are not an independent incident log, and the complaint itself says OpenAI has not answered detailed questions about the full set of events. A later filing, or an OpenAI account given in court, could support, narrow or contest that total.

OpenAI’s statement, as quoted by CNBC, should also be kept in its place. It does not appear in the complaint or the AG’s release. In it, a company spokesperson describes post-incident reports, notifications and tighter safeguards, and says the suit singles OpenAI out for a problem other major labs have also identified. That is OpenAI’s response to the case, not a judicial finding that the allegations are wrong, or that the company’s remedies were enough. The documented event today is that a lawsuit has been filed, that the AG has set out what he wants the court to order, and that OpenAI, through a statement reported in the press, disputes the way the case is framed.

Sources

Related stories

Follow @tsnmediahub on X for breaking AI news as it happens.

Follow on X

Share this story

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

More in this category

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.