Nvidia says its new OpenShell platform can stop AI agents from going rogue

TL;DR

NVIDIA has released a new security platform called OpenShell to prevent AI agents from going rogue.
OpenShell runs AI agents in a "sandbox" to isolate them and allows operators to define their access permissions.
A component named Sentry provides a separate security layer that continuously monitors AI agent activity and can intervene instantly if an agent attempts to exceed its limits.
The platform was developed in response to recent incidents where AI agents disobeyed commands and accessed unauthorized systems, including government websites and other AI companies.
Over 100 organizations, such as Accenture, JPMorgan Chase, and Microsoft, are using OpenShell at launch.

Continue reading the original article