• TCG068: Agents and Identity - Navigating What We Can’t Predict Podcast:Download(130.0MB) |Embed WilliamCollins EyvonneSharp We’ve spent a decade figuring out how to (more or less) securely authenticate humans. • Now AI agents are crashing the party, and identity just got a whole lot more complicated.Today we sit down withDan Moore, Senior Director of CIAM Strategy and Identity Standards at FusionAuth, to explore the collision course between artificial intelligence and identity management. • We dig into why AI agents need to be treated as distinct entity types (not just glorified service accounts), how MCP’s adoption of OAuth 2.1 made agentic workflows enterprise-ready, and the wilderness of authorization that exists behind every MCP server. • Plus, we discuss the difference between disposable and durable code, and why curiosity is your best career investment right now. • Episode Links: Dan Moore’s Bluesky Dan Moore’s Website Blog Posts by Dan Moore Articles about AI and Authentication Is that allowed? • Authentication and Authorization in Model Context Protocol Authenticating AI Agents: The New Authentication Paradigm FusionAuth Explore our podcasts Have feedback for the hosts?
Article Summaries:
- Summary (≈100 words) The episode explores how the rise of AI agents is reshaping identity management. Dan Moore, FusionAuth’s Senior Director of CIAM Strategy, explains that AI agents must be treated as distinct entity types rather than simple service accounts. He highlights how MCP’s adoption of OAuth 2.1 has made agent‑centric workflows enterprise‑ready, while also pointing out the complex authorization layers that exist behind every MCP server. The discussion touches on the trade‑offs between disposable and durable code and concludes with a note that curiosity and adaptability are now key career assets in the evolving AI‑identity landscape.
Sources: