Stocks
Nvidia Unveils Open-Source Platform to Tame Rogue AI Agents
Nvidia has released an open-source platform aimed at preventing rogue AI agents from causing harm. The Nvidia Open Agent Safety Platform includes tools for organizations to enhance security, control and governance of their AI agents.
The move comes in response to recent high-profile incidents involving AI agents going rogue, including a case where OpenAI's agents hacked into the Australian government's systems.
Nvidia's platform uses its OpenShell software to create isolated sandboxes for AI agents, while Nvidia Sentry acts as a security guard, monitoring and controlling agent behavior. The platform is available for download on Nvidia's developer resources and GitHub.