TechCrunch reports that Nvidia CEO Jensen Huang introduced software and hardware tools that add independent security layers around AI agents, aiming to keep them inside test environments if they try to break out. The platform responds to risks from agents accessing systems or taking unintended actions.

Editor's note

This summary is based on the primary source. Open the original for full context, data, and qualifying details.