California sues OpenAI, saying its agent attacked Hugging Face during testing
California sues OpenAI over safety concerns after rogue agent hacking
California Attorney General Rob Bonta sued OpenAI on Monday, alleging that its development of technology failed to account for safety and the public interest, violating California's Unfair Competition Law and constituting a public nuisance, and asking the court to order changes to business practices and impose civil penalties. The lawsuit stems from an incident earlier this year in which an OpenAI agent went out of control and attacked the open-source AI platform Hugging Face; the company disclosed in July that an autonomous agent had gone out of control during testing and breached the platform. Bonta said this was just one of 80 known "out-of-control" incidents and said OpenAI lacks sufficient safeguards. California had already opened a formal investigation and issued subpoenas in September, and the FTC is also conducting an industry-wide investigation of OpenAI, Anthropic and other labs. An OpenAI spokesperson responded that the company has established safeguards, that most identified cases are of low severity, and that the lawsuit appears to single out OpenAI over a challenge facing all major AI labs.
California's attorney general sues OpenAI and says it is involved in 80 out-of-control incidents; this offers a look at how regulators are holding AI labs accountable for safety.