Nvidia Unveils Open-Source AI Security System Amid Rogue Agent Concerns
Nvidia has introduced an open-source AI security system in response to recent reports of rogue agents causing chaos online. The chipmaker is rallying tech companies to use its new tool for AI security, which includes OpenShell and Sentry.
OpenShell is a framework for containing agents as they carry out tasks and isolating their activity in the operating system kernel. It's now entering general release for all users after being first announced at Nvidia's annual GTC Conference in March. Nvidia has also developed Sentry, an isolated security domain for chips that continuously monitors long-running AI agents.
Nvidia is positioning its new open-source framework as part of a broader effort to reduce AI risks. The company launched an industry-wide AI safety coalition last July with more than 120 companies participating, and it's also working on the Shared AI Findings Exchange (SAFE) program. However, some experts have raised questions about Nvidia's influence in setting standards for AI technology.