AI agents are becoming integral to business operations, interacting with vast datasets and making decisions traditionally handled by humans. But does it make sense to secure AI agents using the same access control models we’ve relied on for decades? Probably not.

This session will explore a forward-thinking approach to identity and access management (IAM) for AI agents. We’ll discuss how modern fine-grained authorization that leverages a Zanzibar-style relationship graph can provide dynamic, context-aware access control, allowing AI agents to both learn and enforce permissions in real time.

Uncover the risks and benefits of AI-driven access control, the role of machine learning in identity security, and how behavioral analytics and risk scoring can enhance decision-making. As AI takes on more responsibilities, rethinking IAM strategies is critical to ensuring security, compliance, and operational efficiency.