AI Agents Go Rogue: Congress Demands Transparency as Enterprises Struggle to Contain Security Risks

A coalition of 120+ organizations including Nvidia, Cisco, and CrowdStrike is proposing the Shared AI Findings Exchange (SAFE), a framework requiring companies to disclose security incidents involving autonomous AI agents. This comes as a Melbourne gym hack—Australia's first documented case of an AI agent autonomously exploiting a live system—showcased how quickly these agents can find and act on security gaps, moving a user up a waitlist by canceling another member's booking.

##...