AI Agents Need Continuous Trust, Not Just Authentication

Four major tech companies independently reached the same conclusion at RSAC 2026: traditional zero trust security is insufficient for AI agents. Microsoft, Cisco, CrowdStrike, and Splunk each called for new governance models that go beyond one-time authentication. Cisco's Matt Caulfield warned against letting agents run unchecked after initial login, advocating for continuous verification of every action. Jeetu Patel compared agents to teenagers with no fear of consequence, arguing the industry must shift from access control to action control to contain the blast radius of compromised AI systems.