字号·· | 护眼
The Information

白宫人工智能工作组第二次会议讨论网络威胁与关键基础设施White House AI Task Force Discussed Cyber Threats, Critical Infrastructure at Second Meeting

点「原文对照」整页切到原文,或双击某段只看那段的原文。

白宫人工智能工作组周四举行了第二次会议,一名政府官员表示,会议重点在于落实主要人工智能公司(包括Anthropic和OpenAI)之间达成的协议,审查特朗普政府关于在公开发布前测试前沿人工智能模型的自愿框架,并致力于采取措施强化美国关键基础设施以抵御网络威胁。

The White House’s task force on artificial intelligence held its second meeting on Thursday, according to an administration official, focusing on implementing an accord between major AI companies including Anthropic and OpenAI, reviewing the Trump administration’s voluntary framework for testing frontier AI models before public release and working on measures to harden U.S. critical infrastructure against cyber threats.

该小组被称为“超级智能工作组”,由国家情报总监杰伊·克莱顿牵头,于10月初宣布成立,此前白宫曾与人工智能领域领袖举行会议,与会者包括Anthropic的达里奥·阿莫代伊和英伟达的黄仁勋。在那次会议上,各公司同意遵守一套自愿安全原则,包括模型对齐的内部控制和外部审计。该小组还包括联邦贸易委员会主席安德鲁·弗格森和国防部高级官员埃米尔·迈克尔,后者上个月告诉《The Information》,优先事项之一是与人工智能公司合作,制定应对网络入侵和其他关键事件的协议。

Led by National Intelligence Director Jay Clayton, the group—known as the Super Intelligence Force—was announced in early October following a White House meeting with AI leaders, including Anthropic’s Dario Amodei and Nvidia’s Jensen Huang. At that meeting, the companies agreed to a voluntary set of safety principles, including internal controls on model alignment and external auditing. The group also includes Federal Trade Commission chair Andrew Ferguson and senior Defense Department official Emil Michael, who told The Information last month one priority is working with AI companies to develop protocols for cyber breaches and other critical incidents.

据该官员透露,第二次会议的大部分时间集中讨论了这些人工智能风险,包括国家支持的行为体构成的威胁。包括商务部长霍华德·卢特尼克、国家网络总监肖恩·凯恩克罗斯和科技政策办公室主任迈克尔·克拉齐奥斯在内的多位政府高级官员,通报了政府各部门在模型公开发布前如何对其进行评估。该小组还推进了若干政策建议,包括重新授权一项关于网络安全信息共享的国会法案,以及完成网络事件报告相关规则的制定。

Much of the group’s second meeting focused on those AI risks, including threats posed by state-sponsored actors. According to the official, several other senior administration officials, including Commerce Secretary Howard Lutnick, National Cyber Director Sean Cairncross and Office of Science and Technology Policy director Michael Kratsios, provided an update about how government offices are evaluating models before their public release. The group also advanced several policy recommendations, including reauthorizing a Congressional measure on cybersecurity information sharing and finalizing rulemaking around cyber incident reporting.

此次会议是在围绕黄仁勋的丑闻不断发酵之际召开的。黄仁勋是白宫的常客,也是特朗普的亲密密友,他在周四上午宣布辞职。这位政府官员表示,鉴于该工作组专注于前沿模型能力而非芯片制造商,他们预计英伟达事件不会立即对工作组的工作产生影响。尽管如此,黄仁勋的惊人倒台已在特朗普政府内部引发震动。周五上午,财政部长斯科特·贝森特在CNBC节目中批评卢特尼克和商务部在打击芯片走私问题上行动不够迅速。

The meeting comes amid a mounting scandal surrounding Huang, a frequent visitor at the White House and close confidant of Trump, who announced his resignation Thursday morning. The administration official said they do not expect immediate impact from the Nvidia fallout on the task force’s work, given its focus on frontier model capabilities rather than chipmakers. Still, Huang’s stunning fall has sent shockwaves through the Trump administration. On a Friday morning appearance on CNBC, Treasury Secretary Scott Bessent criticized Lutnick and the Commerce Department for not moving more quickly to address chip smuggling.

阅读媒体原文 ↗