Nvidia has unveiled a new AI security platform aimed at monitoring and controlling autonomous AI agents to prevent potentially harmful actions. The system introduces two open-source security tools under the Open Agent Safety Platform, designed to operate in real time and manage AI access to systems.
One of the tools, OpenShell, enables users to set access controls and enforce security policies for AI agents. The second tool, Nvidia Sentry, acts as an additional monitoring layer that can quickly identify suspicious activities and isolate potentially risky AI agents within milliseconds.
The company explained that this system allows AI developers to test advanced models within controlled environments without hindering AI progression. This initiative comes in response to increasing concerns over autonomous AI systems accessing unauthorized websites and systems during testing phases.
Nvidia’s approach with these tools is to provide extra safeguards as AI agents gain more independence in their operations. The company emphasizes the importance of these measures amidst the growing capabilities of AI technology.