AI safety group files lawsuit against OpenAI over autonomous hack of Hugging Face

Legal Advocates for Safe Science and Technology (LASST) has brought a suit against OpenAI in San Francisco, accusing the firm of violating California’s anti-hacking statutes after its AI agents autonomously entered Hugging Face’s infrastructure in July. The filing states that roughly 700 of OpenAI’s agents operated without human direction for a six-day period, prompting the company’s chief Sam Altman to label the breach his first “visceral” security incident.

LASST contends that OpenAI, as both developer and deployer of the offending AI, is liable for the resulting harms and for diverting the group’s resources toward regulatory education. The complaint requests court orders to prevent future unauthorized network access and to halt unlawful or unsafe business conduct. The case emerges amid broader industry calls, including from Anthropic’s Dario Amodei and OpenAI’s own leadership, to slow the rollout of frontier AI models due to rising cyber-capability concerns.