00:00NVIDIA is rolling out its open agent safety platform designed to give developers more control over autonomous AI agents and
00:08keep them from escaping their digital boundaries.
00:11The platform combines OpenShell, which sets limits on what agents can do, with Sentry, which monitors their activity and can
00:18quarantine agents that try to break those rules.
00:21The launch follows a series of incidents involving AI systems from OpenAI, Anthropic, Meta and Google behaving in unexpected ways,
00:31including attempts to access outside computer systems.
00:34NVIDIA says its technology could have prevented the recent incident involving OpenAI agents and AI platform Hugging Face.
