Security Level Five CEO谈模型权重安全与失控智能体风险
The Information播客第9期请来Security Level Five的CEO聊模型权重防盗和失控智能体,这个话题平时很少被讲到基础设施层面。如果你在管模型部署或关注AI安全,可以听听安全创业公司怎么看“看住模型本身”这件事。
The Information播客AI Deep Dive第9期邀请Security Level Five联合创始人兼CEO Lisa Thiergart,与主持人Rocket Drew对谈,节目于太平洋时间12点上线。
本期核心话题是模型权重(model weights)外泄问题,即攻击者或系统本身如何窃取一个前沿模型的全部参数。Thiergart还讨论了rogue AI agents(失控智能体)的情景,以及为“容纳”高级AI系统而建设安全基础设施所面临的技术挑战。
这场对话聚焦的是AI安全行业中的一个细分方向:当模型本身成为需要被物理和网络层面看管的资产时,企业需要什么样的防护体系。对于关注AI安全、模型部署合规的企业技术决策者,这期节目提供了一个来自安全创业公司一线视角的入门材料。
What would it take to keep advanced AI systems—and their model weights—secure?
For Episode 9 of AI Deep Dive, Lisa Thiergart, co-founder and CEO of Security Level Five, joins Rocket Drew to discuss model exfiltration, rogue AI agents and the technical challenges of building infrastructure capable of containing advanced AI.
🚀 Watch today at 12PM PT / 3PM ET: https://t.co/kYW16CDQIa
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Key entities and technical topics under continuous tracking include AI安全, Security Level Five, The Information, 模型权重.
Source attribution: curated from The Information (@theinformation).