2026 Research Experiences for Teachers (RET) Poster - Computational Pathology and Informatics

0:00 Introducing the PathAssist Summer Project 0:27 How PathAssist Supports Diagnostic Decisions 1:00 Integrating AI Into Pathology Workflow 1:35 Phase One Needs Assessment 2:27 What Users Want From PathAssist 3:05 Building Trust Through Evidence and Uncertainty 3:47 Phase Two Prototype Review 4:17 Designing the PathAssist User Interface 4:33 Prototype Feedback and Improvement Opportunities 5:13 AI Literacy and the Hopeful Skeptic 5:48 Closing Reflections on Responsible Clinical AI2026  Mayo Clinic RET Fellow presents end-of-program poster in collaboration with Mayo Clinic Division of Computational Pathology and Informatics. How can AI support pathologists without replacing clinical judgment? In this presentation, Nick Elias shares his Mayo Clinic RET summer project exploring the user-centered design of PathAssist, an AI-assisted pathology decision-support tool designed to help physicians evaluate ancillary testing options during diagnostic workup. The project examines what pathology residents and fellows need from an AI interface to understand recommendations, evaluate uncertainty and appropriately trust the information being presented. Findings highlight the importance of transparent confidence measures, case-specific evidence, differential diagnoses and clearly ranked test recommendations. The presentation also explores prototype feedback, opportunities to improve the PathAssist interface and a broader lesson for responsible AI education: effective AI literacy requires knowing when to trust a tool, when to question it and when to seek additional evidence. Topics include: • AI-assisted pathology decision support • Communicating uncertainty and model confidence • User-centered clinical interface design • Building appropriate trust in AI recommendations • Responsible and ethical AI use in medicine 0:00 Introducing the PathAssist Summer Project 0:27 How PathAssist Supports Diagnostic Decisions 1:00 Integrating AI Into Pathology Workflow 1:35 Phase One Needs Assessment 2:27 What Users Want From PathAssist 3:05 Building Trust Through Evidence and Uncertainty 3:47 Phase Two Prototype Review 4:17 Designing the PathAssist User Interface 4:33 Prototype Feedback and Improvement Opportunities 5:13 AI Literacy and the Hopeful Skeptic 5:48 Closing Reflections on Responsible Clinical AI