Define identity, authentication, and authorization architectures for AI agents, including user, workload, service, and agent identities; Role-Based Access Control (RBAC); delegated authorization; least-privilege access; multi-tenant isolation; and secure propagation of identity and authorization context across distributed workflows. Partner across engineering, research, product, security, privacy, and compliance organizations to translate evolving AI capabilities, risks, and enterprise requirements into scalable architectures and balance quality with reliability, latency, cost, security, safety, and operational complexity.