LASST Lawsuit Accuses OpenAI of Unsafe AI Development Practices After Hugging Face Hack

After the Hugging Face hack, staff at LASST set aside their normal workload to “plan, coordinate, and participate in regulatory briefings regarding this incident.” The organization has since responded to additional briefing requests that required further work.

According to the complaint, LASST staff spent dozens of working hours addressing what the organization describes as OpenAI’s unsafe development practices.

“Despite the impact on LASST’s other programs, LASST has committed resources to efforts to combat OpenAI’s illegal conduct,” the lawsuit states. “If LASST prevails in this case, it will no longer have to divert resources to combat the illegal and unfair business practices employed by OpenAI in which its AI agents hack third parties during internal evaluations.”

LASST Seeks Injunction Against OpenAI

According to the complaint, the requested injunction would prohibit OpenAI from “accessing or causing unauthorized access to any computer, computer network, or computer system, either by itself or through any artificial intelligence agents it develops, deploys, modifies, or uses.”

The injunction would also prohibit OpenAI from knowingly employing unfair business practices that threaten to cause serious harm to the public.

U.S. lawmakers from both major political parties asked for an answer after the incident. The proposed AI Kill Switch Act, a proposal from OpenAI, would allow U.S. government authorities to order the shutdown of dangerous AI systems.

LASST said new regulations are needed to protect the public from AI risks. However, the organization said California’s current law allows it to rein in AI companies “without waiting for new regulations to catch up with the harms that are happening to businesses and consumers today.”

“We are filing this lawsuit because OpenAI violated the law, and OpenAI must be held accountable,” the group said. “OpenAI and broader frontier AI developers cannot avoid the consequences of their dangerous actions simply by claiming ‘AI did it.’ Autonomous AI agents will continue to hack, steal data, disrupt systems, and violate rights until the courts intervene.”