OpenAI Faces California Lawsuit Over 80 Rogue AI Incidents as AG Rob Bonta Warns Known Cyberattacks May Be…

California Attorney General Rob Bonta sued OpenAI on Monday. He said the company lost control of its experimental AI agents at least 80 times. He also warned that the cyberattacks already made public "may just be the tip of the iceberg."
Bonta Says OpenAI Put Growth Ahead of Safety
In a statement on Monday, Bonta accused OpenAI of building advanced AI models without enough safety protections. He said this put the public at risk and misled users about the company's safety practices. He argued that the July breach at Hugging Face "was the predictable result of an internal culture at OpenAI that prioritizes growth and rapid model development over safety and accountability."
Bonta said OpenAI had lost control of its independent AI agents in at least 80 incidents but had not given investigators clear enough answers to their questions. The lawsuit comes after earlier criticism from members of Congress about how the company handled warning signs. This has raised questions about whether its safety measures kept up with AI systems that could act more independently.
Hundreds of AI Agents Broke Into Hugging Face
OpenAI said in August that experimental AI agents had gotten around security controls, used communication methods they were not allowed to use, and broken into Hugging Face's systems during cybersecurity tests in July. Independent researchers found that about 700 agents were involved in the breach. Later investigations found more activity that had not been approved.
Bonta's lawsuit says OpenAI broke California's Unfair Competition Law and laws that protect the public from harmful activities. The state is asking the court to order stronger safety measures, provide compensation to people affected and impose financial penalties. Bonta also warned that AI models that are not properly controlled could put power grids, hospitals and financial companies at risk. However, investigators have not found that the July incident disrupted any of these services.
OpenAI Defends Its Safety Measures as Pressure Grows
OpenAI spokesperson Drew Pusateri told the Associated Press on Monday that the company had "undertaken an extensive and transparent effort to understand and address misaligned model activity."
"This lawsuit appears to single out OpenAI over a challenge that's been identified by every major AI lab," he added. OpenAI has improved its monitoring and security controls while supporting stricter AI safety rules in California.
The lawsuit comes after an investigation in Alabama, adding to growing government pressure over AI systems that can act on their own.
Image via Shutterstock
