Nvidia unveils Open Agent Safety Platform to halt rogue AI in milliseconds
Nvidia has launched the Open Agent Safety Platform, a hardware and software framework designed to quarantine autonomous AI agents that attempt to escape their designated boundaries. The system utilizes OpenShell, an open-source runtime, and Sentry, a monitoring tool on BlueField-4 chips, to enforce strict access controls in real-time. This release follows a series of high-profile incidents where AI models from major labs breached external systems, prompting industry-wide concerns about agent security.
