In the rapidly evolving digital landscape, AI agents and advanced chatbots are increasingly interacting with sensitive organization  data on our behalf, blurring the lines between human and machine identities. This presentation explores the challenges and potential solutions in managing identities for these non-human entities. We'll explore how OAuth 2.0's concepts of scopes, tokens, and delegated authority could be extended to create a flexible, secure framework for AI agent identity management. This approach could allow for granular control over AI agent permissions while maintaining the principle of least privilege.This  presentation aims to spark discussion on the future of identity management in an AI-driven world and propose practical steps towards securing our digital interactions in this new frontier.