Nvidia releases Open Agent Safety Platform on Monday to prevent AI models from escaping containment after OpenAI, Anthropic, Meta, Google reported recent security breaches; includes OpenShell and Sentry tools with Microsoft, Cisco, Oracle as partners.

Nvidia releases software platform to stop AI agents from misbehaving

Nvidia CEO Jensen Huang speaks during the keynote address at Salesforce's Dreamforce conference at the Moscone Center on Sept. 15, 2026 in San Francisco, California. Benjamin Fanjoy | Getty ImagesNvidia is rolling out a new software platform to allow AI developers to set safeguards for agents and prevent them from breaking out of containment.The release on Monday of Nvidia's Open Agent Safety Platform comes after companies including OpenAI, Anthropic, Meta, and Google disclosed recent incidents ...