Nvidia Rolls Out Open Agent Safety Platform for Secure AI Agents
Nvidia has launched the Nvidia Open Agent Safety Platform, a software and hardware combination designed to keep AI agents contained. The platform is an open-sourced reference design with a mix of partners.
The safety platform's key components include Nvidia OpenShell, which enables developers to set limits on what agents can do, and Nvidia Sentry, which runs outside the agent environment for added security.
Nvidia CEO Jensen Huang said AI's potential will only be realized if we solve AI safety. He emphasized that deterministic rules need to govern probabilistic agents, as agents can drift when instructions are ambiguous.