California Sues OpenAI Over 'Hugging Face Hack' ... "Involved in at Least 80 Attacks"
California Attorney General Rob Bonta
Sues OpenAI for Violating the Unfair Competition Law
"Tested Dangerous AI Models and Failed to Follow Safety Procedures"
California prosecutors have sued OpenAI, alleging that it failed to prevent cyberattacks involving its artificial intelligence (AI) models. They claim that at least 80 cyberattacks occurred because OpenAI prioritized development speed over safety.
According to Bloomberg on Nov. 16 (local time), California Attorney General Rob Bonta filed a lawsuit against OpenAI that day in state court in San Francisco, accusing the company of violating the state's Unfair Competition Law and creating a public nuisance, among other claims.
The state alleges that OpenAI tested dangerous AI models without appropriate oversight, failed to follow safety procedures, and made statements that misled the public. This is the first time a U.S. state government has sought to hold a company legally responsible for cyberattacks involving an AI model.
The incident at issue was the unauthorized intrusion into the external open-source AI platform Hugging Face in July. In July, OpenAI's AI models, including GPT-5.6 Sol, reportedly accessed the external internet and broke into Hugging Face servers without authorization while undergoing performance tests. In the course of carrying out assigned tasks, the AI accessed external systems beyond the developers' control.
OpenAI said it was strengthening new safeguards to prevent unauthorized actions by its AI models. However, Attorney General Bonta criticized the company, saying, "The Hugging Face attack was a foreseeable consequence of OpenAI's internal culture, which prioritizes growth and rapid model development over safety and accountability."
The state particularly alleges that the Hugging Face incident was just one of at least 80 cases of unauthorized cyberattacks involving OpenAI's AI agents. Bonta emphasized that OpenAI's AI testing activities could threaten any computer or network connected to the internet, saying they "heightened cybersecurity risks and endangered the safety, health, and welfare of society."
California has therefore asked the court to impose a civil penalty of $2,500 (3.34 million won) for each violation of laws on unfair business practices, recover any ill-gotten gains, and issue an injunction ordering OpenAI to stop its unlawful conduct.
Attention is also focused on whether the lawsuit will prompt similar action by other states and against other companies. Anthropic, Meta, Google, and other AI companies have previously disclosed that their models had at times become uncontrollable.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
OpenAI told Bloomberg, "Although unauthorized AI model behavior has been observed across all major AI research labs, this lawsuit singles out OpenAI. Since the Hugging Face incident, we have made extensive and transparent efforts to understand and address the faulty AI model."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.