Nvidia CEO Jensen Huang unveiled the Open Agent Safety Platform on September 28, a two-part security system for autonomous AI agents consisting of OpenShell, an open-source runtime under the Apache 2.0 license, and Sentry, a hardware-level monitoring layer. The system works by isolating AI agents in sandboxed environments similar to browser tabs, allowing each agent to run independently while maintaining audit trails and pushing policy updates without interruption. Over 100 enterprises, including Adobe, Salesforce, SAP, Microsoft, Red Hat, and Anthropic, have already integrated the platform, which targets four domains: coding, research, cybersecurity, and operations. Huang emphasized that AI’s extraordinary potential for society will only be realized if AI safety challenges are addressed.
Source: Read the original article
