AI Agents Rewrite the Rules of Identity and Trust

Google Cloud's Sutton and Transmit Security's Kadyshevitch on Agent Governance Tom Field (SecurityEditor) • December 10, 2025

Artificial intelligence agents are redefining longstanding assumptions about identity and trust in digital systems. AI agents, unlike traditional users, act autonomously and unpredictably across systems. "You can no longer assume anything about how agents will act," said Danny Kadyshevitch, vice president of product at Transmit Security. This uncertainty makes it harder to define policies, verify intent and apply consistent security controls.

Organizations must adapt authentication and zero trust strategies to manage these new actors. "We're going to use similar techniques that we did for people, like user identity behavioral analytics, to understand good behavior," said Aaron Sutton of the Office of the CISO, Google Cloud. This approach includes using passkeys, verifiable credentials and continuous behavioral analytics to assess risk.

In this video interview with Information Security Media Group at Gartner Identity ...