Why You Can’t Trust AI Agents, with SecureAuth's Geoffrey Mattson

On this episode of ID Talk, Alex Perala and Tony Bitzionis speak with Geoffrey Mattson, CEO of SecureAuth, about the growing security risks surrounding agentic AI. As enterprises rush to deploy AI agents for coding, workflow automation, and decision-making, Mattson argues that identity security has gone from being an “NPC” in cybersecurity to having “main character energy.” 

From executives accidentally exposing sensitive data through vibe coding to autonomous agents making decisions no human intended, the conversation explores why companies need guardrails around what AI agents can do, not just trust in the models themselves.