Nvidia Adds Hardware-Level Security Layer to Keep Rogue AI Agents Contained

Nvidia CEO Jensen Huang has introduced the Nvidia Open Agent Safety Platform, a software and hardware toolkit that adds independent security layers around AI agents to keep them within their test environments even if they try to break out.