NVIDIA has introduced its Open Agent Safety Platform, an open software platform and reference system design aimed at putting enforceable security boundaries around AI agents from testing through deployment. Rather than relying only on safeguards inside an AI model or agent framework, NVIDIA is extending security controls into the runtime, compute infrastructure, and hardware that […]
English