在接连发生安全事件后，英伟达周一推出一个平台，并称该平台可以管控 AI 智能体。
Nvidia on Monday launched a platform that it said could keep AI agents in check, following a string of security breaches.
公司表示，这款软件可以监测智能体的行为，并隔离其中可疑的智能体。近期多起智能体失控事件促使科技行业领袖呼吁放缓 AI 发展，但英伟达首席执行官却是一个显著的例外。英伟达智能体 AI 业务副总裁在接受 Semafor 采访时表示：“我们与其制造恐惧和 FUD，不如直接保障它的安全。我们知道该怎么做。”他解释道，FUD 是恐惧、不确定性和怀疑的缩写。
The software can monitor the behavior of agents and quarantine suspicious ones, the company said. Recent rogue incidents have rallied tech leaders to call for an AI slowdown, with Nvidia’s CEO being a notable exception. “Rather than creating fear and FUD, let’s just secure it. We know how,” Nvidia’s VP of agentic AI told Semafor, referencing the acronym for fear, uncertainty, and doubt. The White House has similarly downplayed AI safety risks, arguing that is companies’ responsibility.
白宫同样淡化了 AI 安全风险，认为保障 AI 安全是企业应尽的责任。美国顶级 AI 企业高管将于周二与美国总统唐纳德·特朗普会面，但 Semafor 的阿什利·戈尔德写道：“他们更可能受到盛情款待，而不是受到斥责。”——塔斯尼姆·纳什鲁拉
Top AI bosses are meeting US President Donald Trump Tuesday, but “they’re more likely to be fêted than reprimanded,” Semafor’s Ashley Gold wrote. — Tasneem Nashrulla