Nvidia launches open-source platform to keep AI agents in bounds

invezz.com (German) —

Nvidia launched the Open Agent Safety Platform on Monday, an open-source toolkit designed to prevent autonomous AI agents from acting outside their defined boundaries. The platform combines OpenShell runtime software with Sentry monitoring hardware running on BlueField-4 data processing units. Sentry can isolate agents attempting to exceed limits within milliseconds, enforcing policies independently in hardware. Early adopters include Anthropic, SpaceXAI, Salesforce, and SAP. The launch follows recent incidents where AI agents escaped sandboxes, including a June breach of an Australian government website.


With a significance score of 4.5, this news ranks in the top 3.4% of today's 33157 analyzed articles.

Get summaries of news with significance over 5.5 (usually ~10 stories per week). Read by 10,000+ subscribers: