California Attorney General Rob Bonta announced a lawsuit Monday against OpenAI, saying the California-based artificial intelligence pioneer has recklessly allowed its AI tools to run amok, including the so-called Hugging Face cyberattack.
Mr. Bonta said OpenAI admitted to his office that the firm has “lost control” of autonomous AI agents more than 80 times so far.
But the California-based firm has rebuffed the attorney general’s attempts to investigate, Mr. Bonta said in the new lawsuit, filed in state court in San Francisco.
People were interested in these podcasts
It charges the firm with unfair business practices and declares the firm’s activities an “ongoing public nuisance.”
“The Hugging Face cyberattack was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability,” Mr. Bonta said in the lawsuit.
Hugging Face is another AI platform.
AI tools developed by OpenAI conducted a cyberattack on Hugging Face in July.
OpenAI said the AI agents went rogue, escaping the carefully designed testing environment set up to contain them.
But Mr. Bonta said OpenAI should have seen the signs beforehand and moved to shut things down.
And he said the threat remains.
“While OpenAI has admitted that it lost control of its autonomous AI agents at least 80 times during testing, it has failed to answer the detailed questions our office has asked about these incidents,” Mr. Bonta’s office said.
The Washington Times has sought comment from OpenAI.

Please read our comment policy before commenting.