
Oct 31, 2026 · 25 min
AI agents expose identity governance’s accountability gap
Who Is Responsible When AI Agents Make Decisions | Matt Mills x Shekhar Natarajan
As autonomous systems gain access to enterprise resources, organizations must decide who authorizes their actions and who remains accountable when they go wrong.
- 1Identity governance now covers machine identities, entitlements, and actions—not just employee access and compliance.
- 2AI agents challenge deterministic authorization because their behavior can emerge from complex models rather than fixed rules.
- 3Human accountability, real-time authorization, policy engines, and trust structures are becoming essential as agents enter enterprise operations.
Don't miss
Matt Mills argues that every AI agent should currently have a human in the loop who remains ethically and morally accountable for its actions.
The brief
Matt Mills and Shekhar Natarajan frame identity governance as more than compliance: it now determines who—and what—can access enterprise systems and which actions they may take.
SailPoint’s work with Entro reflects a widening security problem: organizations must discover machine identities, orphaned entitlements, and ownership gaps as non-human identities multiply.
Traditional access controls assume deterministic behavior, but AI agents act through complex models whose emergent decisions can outpace existing identity and authorization frameworks.
The episode’s central tension is accountability: until stronger structures exist, every agent should have a human in the loop who remains ethically and morally responsible.
The proposed answer combines real-time authorization, policy engines, regulation, and trust structures—while acknowledging that organizations often miss harmful behavior amid complexity.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.