字号·· | 护眼
彭博社

白宫敦促放宽责任规定以鼓励AI披露White House Urges Easing Liability to Encourage AI Disclosure

点「原文对照」整页切到原文,或双击某段只看那段的原文。

周四在白宫的一次会议上,唐纳德·特朗普总统人工智能特别工作组的官员建议各机构放宽网络责任政策,以激励人工智能公司报告重大事件。

At a White House meeting on Thursday, officials from President Donald Trump’s artificial intelligence task force recommended that agencies ease their cyber liability policies in an effort to incentivize AI companies to report critical incidents.

据一位因会议未公开而要求匿名的政府官员透露,这些官员告诉各机构,以往针对网络事件采取监管行动的标准可能已经过时,并对重要的信息共享产生了寒蝉效应。

The officials advised agencies that previous standards for taking regulatory action in response to cyber episodes may be outdated and have a chilling effect on important information sharing, according to an administration official who asked not to be named because the meeting was not public.

这位官员补充说,各机构被敦促在未来12个月、并可能更长时间内限制监管责任。

Agencies were urged to place limits on regulatory liability for the next 12 months and possibly longer, the official added.

向人工智能开发者提供责任方面的“胡萝卜”而非“大棒”,这一转变与早先的公开声明相反。早先的声明显示出愿意援引现行法律来惩罚制造不安全产品的公司。就在上个月,同一个AI特别工作组还警告开发者,他们必须报告并解决安全事件,否则将面临潜在的、未具体说明的后果。

The move to offer AI developers a liability carrot, rather than a stick, is a turnaround from earlier public statements that showed a willingness to invoke current law to punish companies that create unsafe products. Just last month, the same AI task force warned developers that they are required to report and resolve security incidents or face potential unspecified consequences.

此次会议此前由Semafor报道。会议召开之际,正值英伟达公司因其首席执行官黄仁勋的婚外情而将其免职,其中一段婚外情对象是一名被怀疑为中国从事间谍活动的女子。这一丑闻以及黄仁勋与唐纳德·特朗普总统和政府高层官员的密切关系引发了严密审视,尤其是对出席了白宫会议的商务部长霍华德·卢特尼克。

The gathering, which was reported earlier by Semafor, took place in the shadow of Nvidia Corp.’s ouster of its chief executive officer, Jensen Huang, over affairs he had, including one with a woman suspected of spying for China. The scandal and Huang’s close relationship with President Donald Trump and top administration officials have generated intense scrutiny, particularly of Commerce Secretary Commerce Secretary Howard Lutnick, who attended the White House meeting.

特朗普于9月底组建了这个名为“超级智能部队”的小组。此前,在一连串失控模型引发的安全事件之后,公众对灾难性风险的担忧日益加剧,使政府难以忽视。特朗普还与人工智能行业最知名开发者中的领导人签署了一项协议,包括英伟达、Anthropic PBC、OpenAI、谷歌母公司Alphabet Inc.和Meta Platforms Inc.。

Trump formed the group — known as the Super Intelligence Force — at the end of September, after growing public concern about catastrophic risks became too hard for the administration to ignore following a spate of breaches by rogue models. He also signed an accord with leaders of the AI industry’s most prominent developers including Nvidia, Anthropic PBC, OpenAI, Google parent Alphabet Inc. and Meta Platforms Inc.

这项不具约束力的协议支持使用外部审计,并支持针对网络安全、生物安全、化学威胁以及AI模型意外行为采取安全防范措施。特朗普称其具有“道德约束力”。几天后,美国联邦贸易委员会就产品安全方面的担忧,提出了对AI公司采取法律行动的可能性。

The nonbinding agreement endorsed the use of outside auditors and safety precautions for cybersecurity, biosecurity, chemical threats and unintended actions by AI models. Trump called it “morally binding.”Days later, the Federal Trade Commission teed up the threat of legal action against AI companies related to product safety concerns.

据这名官员称,在周四的会议上,特别工作组主席、国家情报总监杰伊·克莱顿表示,未来可能会有更多私营公司加入该协议。克莱顿还说,他将继续与协议签署方合作,制定共同的行业标准。

During Thursday’s meeting, Jay Clayton, the force’s chairman and the director of national intelligence, said more private companies could be added to the accord in the future, according to the official. Clayton added that he was continuing to work with the accord’s signers on shared industry standards.

这名人士补充说,五角大楼副部长、特别工作组副主席埃米尔·迈克尔也谈到了他与AI公司在事件报告规程方面的工作。出席会议的还有联邦贸易委员会主席安德鲁·弗格森,他同时也是特别工作组成员,以及科学和技术政策办公室主任迈克尔·克拉齐奥斯和国家网络主任肖恩·凯恩克罗斯。

Emil Michael, a Pentagon under secretary and the task force’s vice chair, also spoke about his work with AI companies on an incident reporting protocols, the person added. Also in attendance were Federal Trade Commission Chair Andrew Ferguson, who also serves on the task force, Office of Science and Technology Policy Director Michael Kratsios and National Cyber Director Sean Cairncross.

卢特尼克、凯恩克罗斯和克拉齐奥斯还讨论了AI模型的发布前测试,这项测试依据6月的一项自愿性行政命令进行。这名官员说,国家安全局以及商务部下属的超级智能创新与标准促进中心都在开展这项测试。

Lutnick, Cairncross and Kratsios also discussed pre-release testing of AI models, which is carried out based on a voluntary June executive order. The National Security Agency and the Center for Advancing Innovation and Standards for Super Intelligence within the Commerce Department are both carrying out the testing, the official said.

特别工作组副主席、人事管理办公室主任斯科特·库珀和凯恩克罗斯谈到了他们为保护关键基础设施和政府系统免受AI驱动的网络攻击所做的工作。同一天,财政部长贝森特在保守派智库发表演讲时说,法国一起针对冰箱系统和血库的网络攻击是由中国的AI模型驱动的。

SI Force Vice Chair Scott Kupor, who is also the director of the Office of Personnel Management, and Cairncross spoke about their efforts to protect critical infrastructure and government systems from AI-powered cyber attacks. The same day, Treasury Secretary Bessent said during a speech at a conservative think tank that a cyber attack on refrigerator systems and a blood bank in France was powered by Chinese AI models.

阅读媒体原文 ↗