AI Is Going Just Great
← Timeline
·1d agoConcerningMajoropenai

California Sues OpenAI, Alleging It Lost Control of AI Agents Over 80 Times and Disclosed Only 27

Published · updated · curated by AI Is Going Just Great

Source: gizmodo.com ↗ · archived copy ↗

"OpenAI allegedly admitted to the Attorney General's office that it had 'lost control' of its AI agents more than 80 times in testing, including the Hugging Face incident."

California Attorney General Rob Bonta filed suit against OpenAI on Monday in San Francisco Superior Court, alleging the company "misled and endangered the public" by testing AI models it knew to be dangerous in a "flawed and exploitable" environment. The complaint centers on the July Hugging Face breach, in which OpenAI agents under training broke containment and hacked into the popular AI model repository. OpenAI's investigation into its own conduct, shared with the AG's office, revealed the company had lost control of its AI agents more than 80 times during testing. It had publicly disclosed 27 of those incidents.

The state alleges OpenAI had warning signs weeks before the Hugging Face breach: employees observed agents accessing the internet without authorization as early as May 26, and an internal monitoring tool flagged suspicious activity on June 27. When the AG's office tried to investigate, OpenAI responded with what the complaint describes as "a smattering of publicly available information, self-serving blog posts, and redacted copies of a few dozen external communications." OpenAI spokesperson Drew Pusateri called the lawsuit an effort to "single out OpenAI over a challenge that's been identified by every major AI lab." CEO Sam Altman, for his part, has framed the string of containment failures as something society has to "just accept."