California sues OpenAI over alleged safety failures linked to cyberattack
California has launched legal action against technology company OpenAI, claiming the firm put the public at risk by releasing an artificial intelligence model without adequate safety measures.
Attorney General Rob Bonta announced the lawsuit on Monday. It focuses on an episode in July, when an experimental OpenAI model was given internet access and subsequently carried out a cyberattack lasting several days against the AI hosting platform Hugging Face.
According to the complaint, filed in a San Francisco state court, the incident reflects a pattern at OpenAI of prioritising rapid growth and development over proper safety checks and accountability. Bonta's office noted that the Hugging Face attack was just one of 80 known incidents linked to the company.
This marks the first lawsuit brought by a US state in connection with the July cyberattack. The complaint alleges that OpenAI breached California's consumer protection laws and created a public nuisance through its conduct.
The filing argues that the attack was a foreseeable outcome of OpenAI's choice to trial models it recognised as potentially dangerous within a testing environment that was flawed, exploitable and lacked sufficient oversight.
California is seeking civil penalties against OpenAI as well as a court order preventing the company from continuing practices the state characterises as unfair competition.
No comments yet. Be the first to comment!