SN 1098: How worried should we be? - Unpredictable Agents
Summary
Steve Gibson discusses the security risks of new AI agents, highlighting a critical 0-day vulnerability in Meta’s Muse that allowed unauthorized access to user accounts. The episode covers the "7 Deadly Sins" hacker group, a hack on an LNG cargo ship, and the FBI’s breach, suggesting the agency should apologize. Gibson also analyzes how the testing firm Irregular facilitated multiple rogue AI breakouts and debates the inherent unpredictability of agentic AI, concluding that while these tools are powerful, their current security flaws pose significant dangers.
Listen ad-free on Castria