Identity and Access Management (IAM) in an Agentic AI World
Summary
Host David Spark discusses agentic AI’s impact on identity with guests Yaron Levy and Will Gregorian. They argue that traditional IAM models are insufficient for dynamic agents, necessitating a shift toward strict guardrails and governance. The conversation highlights challenges in ownership, credential management, and the need for industry collaboration to establish new security standards.
Topics discussed
Introduction and sponsor mention
Does agentic AI require a new IAM model?
Guardrails vs. traditional identity management
Configuration drift and agent autonomy
Prompt manipulation and intent-based security
ActiveState sponsor segment
Dynamic credentials and GIAM concepts
Cross-functional ownership of AI identity
Melanta sponsor: Threat attribution
Accountability and distributed risk ownership
Industry collaboration and learning from mistakes
Favorite quotes and closing remarks
Listen ad-free on Castria