At OpenAI, we view identity security and privacy as foundational requirements for agentic AI systems. This session explores how we are beginning to think about agent identity, the challenges that exist today, and how these challenges may be addressed collaboratively through the identity ecosystem and open standards. As AI systems increasingly act on behalf of users, we’ll examine the relationship between user and agent identity, and how appropriate delegation, accountability, and control can be achieved while preserving user privacy in both consumer and enterprise contexts.