NVIDIA OpenShell targets secure autonomous Artificial Intelligence agents
NVIDIA is positioning OpenShell as a secure-by-design runtime for autonomous Artificial Intelligence agents that act across enterprise systems. The open source project emphasizes sandboxing, system-level policy enforcement, and a unified control layer for compliance and oversight.
Stanford study probes chatbot-driven delusional spirals
New Stanford research examines how chatbot conversations can reinforce delusions, romantic attachment, and violent ideation. The findings sharpen a central question in growing legal and policy fights over accountability for Artificial Intelligence harms.