OpenAI 正面临来自加利福尼亚州总检察长办公室的新诉讼,指控该公司的训练实践和薄弱的网络安全措施对公众构成持续威胁。
本案的核心是2026年7月发生的 Hugging Face 数据泄露事件,在该事件中,OpenAI 的一个预发布模型在突破双方安全系统以完成训练基准测试时发生了失控。在关于该事件的官方报告中,OpenAI 称其为“一个涉及罕见且意外事件叠加的异常场景”。
加州的这起诉讼是政府监管机构首次尝试让一家 AI 实验室就失控智能体(rogue agent)事件承担法律责任,这对 OpenAI 的竞争对手及其客户都具有重大影响。
起诉书围绕 Hugging Face 事件展开,但也引用了涉及 RubyGems、DseWiki、维基媒体基金会以及澳大利亚政府网站的类似事件作为证据,表明这种底层行为仍在持续发生。
起诉书中还指控,另有数十起类似的未披露事件已经发生。
“OpenAI 最近向数十个其他第三方发送了通知,这些第三方的网站或内部系统已被 OpenAI 的智能体渗透。”起诉书写道,“在某些情况下,OpenAI 的智能体能够访问敏感的个人信息,例如包含姓名、出生日期、地址等的员工记录。绝大多数此类事件既未被 OpenAI 披露,也未被受影响的第三方公开讨论。”
具体的指控是,OpenAI 持续的测试实践构成了公共滋扰(public nuisance),理由是缺乏安全措施给公众带来了持续的风险。一项次要的指控则基于营销不安全产品而构成不公平竞争。如果诉讼成功,OpenAI 将面临禁令,“以消除持续的公共滋扰”,这将赋予法院广泛的自由裁量权。
这是第一起专门针对 OpenAI 因失控智能体事件提起的监管案件,尽管其他州也曾考虑过危害公共安全的指控。今年六月,佛罗里达州总检察长起诉该公司,指控其对用户造成有害影响,尤其是对儿童和青少年。八月,阿拉巴马州总检察长启动了一项调查,以调查与此次泄露事件相关的欺骗性做法,但尚未提起诉讼。
加州总检察长 Rob Bonta 在向媒体发表的声明中将此案描述为在面对强大新技术时维护公共安全的努力。
“OpenAI 的行为加剧了现实世界的网络安全风险,威胁到社会的安全、健康和福祉。该公司明知故犯地测试危险产品,却缺乏适当的监督或安全程序,并误导了公众。”Bonta 在声明中说,“法律很简单:在没有适当保障措施的情况下,不要做可能伤害人们的事情。”
OpenAI 在一份声明中强调,该公司在 Hugging Face 事件后投入了大量精力进行调查并改进其系统。“这项诉讼似乎单独针对 OpenAI,而这一挑战是所有主要 AI 实验室都已识别出的问题。”OpenAI 表示,“我们高度重视对齐问题,并且我们已经——并将继续——与总检察长办公室建设性地合作,分享信息,并为整个行业推进强大、实用的安全标准。”
更新时间:太平洋时间下午 12:23:添加了 OpenAI 的声明。
当您通过文章中的链接购买时,我们可能会赚取少量佣金。这不会影响我们的编辑独立性。
Russell Brandom 自 2012 年起一直报道科技行业,重点关注平台政策和新兴技术。他曾在 The Verge 和 Rest of World 工作,并为 Wired、The Awl 和麻省理工学院的《Technology Review》撰稿。您可以通过 [email protected] 或在 Signal 上联系他,号码为 412-401-5489。
查看简介
查看原文
OpenAI is facing a new lawsuit from the California Attorney General’s office, alleging that the company’s training practices and lax security present an ongoing danger to the public.
At the center of the case is the July 2026 Hugging Face compromise, in which saw a pre-release model within OpenAI compromise, both companies’ security systems in order to complete a training benchmark. In its official report on the incident, OpenAI called it “an outlier scenario involving a rare and unexpected confluence of events.”
The California lawsuit represents one of the first attempts by a government regulator to hold an AI lab legally accountable for a rogue agent incident, with significant implications for both OpenAI’s competitors and its clients.
The complaint centers around the Hugging Face incident, but also cites similar incidents involving RubyGems, DseWiki, the Wikimedia Foundation and Australian government websites as evidence that the underlying behavior is ongoing.
The complaint also alleges that dozens of other similar incidents have occurred without being disclosed.
“OpenAI has recently sent notices to dozens of additional third parties whose websites or internal systems have been infiltrated by OpenAI agents,” the complaint reads. “In some cases, OpenAI agents were able to access sensitive personally identifiable information, such as employee records containing names, birth dates, addresses, and more. The vast majority of these incidents have not been disclosed or discussed publicly by either OpenAI or by the affected third parties.”
The specific allegation is that OpenAI’s ongoing testing practices represent a public nuisance, on the grounds that the lack of safety controls present an ongoing risk to the public. A secondary claim alleges unfair competition as a result of marketing an unsafe product. If the lawsuit succeeds, OpenAI would face an injunction “to abate the ongoing public nuisance,” leaving broad discretion in the hands of the court.
It’s the first regulatory case targeting OpenAI specifically over rogue agent incidents, although public endangerment claims have been considered by other states. In June, the Florida attorney general sued the company over allegedly harmful effects on users, particularly children and teenagers. In August, the Alabama attorney general launched a probe to investigate deceptive practices related to the breach, but no case has been filed.
In a statement to the press, the California Attorney General Rob Bonta presented the case as an effort to maintain public safety in the face of a powerful new technology.
“OpenAI’s actions escalated real-word cybersecurity risks, threatening the safety, health, and well-being of society. The company knowingly tested a dangerous product without proper oversight or safety procedures and misled the public about it,” Bonta said in the statement. “The law is simple: Don’t do things that can hurt people without proper safeguards in place.”
In a statement, OpenAI emphasized the company’s extensive effort to investigate the Hugging Face incident and improve its systems in the aftermath. “This lawsuit appears to single out OpenAI over a challenge that’s been identified by every major AI lab,” OpenAI said. “We take misalignment seriously, and we have engaged — and will continue to engage — constructively with the Attorney General’s office to share information and advance strong, practical safety standards for the entire industry.”
Updated 12:23pm PT: Added OpenAI statement.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
Russell Brandom has been covering the tech industry since 2012, with a focus on platform policy and emerging technologies. He previously worked at The Verge and Rest of World, and has written for Wired, The Awl and MIT’s Technology Review. He can be reached at [email protected] or on Signal at 412-401-5489.
View Bio
原文版权归 TechCrunch · AI 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。