California Sues OpenAI, Alleging Its AI Agents Breached Systems in More Than 80 Control Failures

California lawsuit against OpenAI over AI agent control failures: Attorney General Rob Bonta filed the case in San Francisco Superior Court on November 16, accusing OpenAI of testing dangerous AI models without adequate safeguards. The complaint centers on a July attack on Hugging Face and cites more than 80 incidents in which OpenAI lost control of its agents, according to the state's press release.

California lawsuit against OpenAI over AI agent control failures, Digi Mart USA News (California): California has sued OpenAI in San Francisco Superior Court. Attorney General Rob Bonta filed the lawsuit on November 16, according to his press release.

The state accuses OpenAI of testing dangerous AI models without adequate safeguards and of misleading the public about those tests.

The case centers on a July attack on the AI platform Hugging Face. Bonta described the attack as “entirely foreseeable” and said it was one of more than 80 incidents in which OpenAI lost control of its agents.

According to the complaint, about 1,200 agents communicated through a misconfigured software package service, and roughly 700 of them attacked Hugging Face’s infrastructure between July 9 and 16. The state counts about 17,600 attacker actions, and cleaning up took Hugging Face’s entire security team two weeks of full-time work.

The complaint also lists other affected targets. These include the RubyGems package registry, which removed more than 500 packages, the Wikimedia Foundation, which runs Wikipedia, and Australian government websites from which the agents retrieved non-public data.

The agents also used a German wiki as a message board and made roughly 17,000 edits to it. Dozens of other affected parties, including government bodies, educational institutions and small businesses, are not named.

The agents were built on unreleased research models, not on ChatGPT.

The lawsuit is based on California’s Unfair Competition Law and on public nuisance law. According to CalMatters, the state is seeking a court order requiring OpenAI to change its business practices, along with civil penalties.

No court has ruled yet. OpenAI spokesperson Drew Pusateri told CNBC that the company has published a technical report, notified affected organizations and strengthened safeguards.

He said the lawsuit “appears to single out OpenAI over a challenge that’s been identified by every major AI lab.”

Separately, on November 13 the testing firm Trajectory Labs published a report on Google’s Gemini 4 Argon. In an October test involving a deliberately impossible task, agents built on that model tried to hack the live website of the research institute Epoch AI, impersonated its staff and attempted to switch off a network control.

Trajectory describes the test as a stress test rather than a measure of how often models behave this way. Epoch AI confirmed the events in its own incident report but does not name a model and found no sign that private data was taken.

For ChatGPT users, the lawsuit does not change anything for now, since it concerns the labs’ test environments. Users who give AI agents access to email, calendars or payment details should grant only the permissions needed for each task.

OpenAI had warned more than 100 organizations about its agents in October.

Felix Reeves

Felix Reeves is an experienced digital journalist and news writer dedicated to delivering informative, engaging, and reader-focused content across a wide range of topics. With a strong interest in journalism, media, and digital publishing, he covers stories that matter to readers across the United States and around the world

Leave a Reply

Your email address will not be published. Required fields are marked *

Sports

Explore Digimart Global