Open AI is being sued by California which claims the company misled and endangered the public over its development of advanced AI models.
According to the lawsuit, Open AI knowingly developed and tested artificial intelligence without proper guardrails and monitoring that led to an infamous hacking incident in July. Around 700 of Open AI's autonomous agents escaped its testing sandbox to launch two days worth of cyber attacks against the AI platform, Hugging Face.
State Attorney General Rob Bonta said without safeguards, these models pose a danger to anything connected to the internet.
“They can intrude into, degrade or shut down the services people rely on every day, including our power grids," Bonta added, "our hospital systems, our financial institutions, and other critical infrastructure.”
Bonta said the Hugging Face incident is one of 80 cyber incidents in which the company has admitted to losing control and agents gained access to third party networks. The lawsuit claims the company created a public nuisance and a risk to public safety. Open AI is also accused of engaging in unfair or fraudulent business practices.