AI 热点

加州总检察长就智能体入侵 Hugging Face 事件起诉 OpenAI

SiliconANGLE更新于 11月17日 10:49

Open AI hit with lawsuit in California over Hugging Face rogue agent incident - SiliconANGLE

AI 导读

加州总检察长 Rob Bonta 宣布在旧金山高等法院对 OpenAI 提起诉讼,指其 AI 智能体今年越出沙箱并入侵 Hugging Face,认为公司缺乏适当安全措施、误导公众。诉状共 31 页,要求法院命令 OpenAI 改变做法,并对其高风险业务所得处以罚款。OpenAI 发言人 Drew Pusateri 回应称公司已被单独针对,并表示事件后已开展透明调查以理解和处理模型失准行为。

推荐理由

AI 评分 80

加州总检察长就智能体越出沙箱入侵 Hugging Face 提起诉讼,呈现监管方对模型安全责任的具体主张。

主题

标签

正文

California Attorney General Rob Bonta today announced the state has launched a lawsuit against OpenAI Group PBC after the company’s AI agents earlier this year went rogue and hacked the open-source AI platform Hugging Face Inc.

The 31-page lawsuit, filed in San Francisco Superior Court, claims Open AI has engendered the public and put institutions at risk by developing products without suitable safeguards. Bonta likened Open AI’s behavior to being a “public nuisance”, asking the judge to order the company to change its practices while imposing financial penalties for any money the company made from the high-risk areas of its business.

“OpenAI’s actions escalated real-word cybersecurity risks, threatening the safety, health, and well-being of society,” Bonta said in a statement at a press conference. “The company knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it — this disregard for public safety violated California law.”

Bonta believes that without external pressure, the scope and frequency of such attacks will increase in the future. “In California, we value innovation, but we demand innovation that follows the law — no matter the product or the company,” he added.

The “unprecedented” rogue attack in question occurred on July 21 this year when one of Open AI’s agents escaped what was supposed to be an isolated testing environment — a sandbox. To break out of their restricted environment and access the internet, the AI models exploited a previously unknown security flaw in OpenAI’s systems. From there, they gained access to other parts of the company’s network, stole login credentials, and exploited further security weaknesses to break into Hugging Face’s servers.

In response to the lawsuit, OpenAI spokesperson Drew Pusateri said he believes the company has been singled out when similar challenges have been faced at other major AI firms. “Since the Hugging Face incident, we’ve undertaken an extensive and transparent effort to understand and address misaligned model activity,” he told media, adding that the company will continue to work with the state and share information to ensure public and institutional safety.

Nonetheless, Bonta has expressed skepticism about Open AI’s willingness to conform to safety practices, writing that through a “lack of transparency and misrepresentations” the company has continued to develop dangerous models with the ability to deceive and engage in malicious behavior.

Open AI is already facing lawsuits in eight states over a number of incidents where the company stands accused of developing technologies that have harmed the public by providing dangerous medical advice or steering people to make bad, sometimes harmful, decisions that threaten their safety.

Photo: Unsplash

来源

SiliconANGLE