This abstract was part of the 2026 UCSF AI and Education Symposium
Proposal: Artificial intelligence has the potential to fundamentally transform medical education by upgrading passive learning to learning with active, high-fidelity simulation. Currently, a large degree of medical education relies on multiple choice questions with pre-packaged histories and the correct answer listed somewhere between choice “A” and “D”, a luxury that doesn’t exist in real life. We propose a show-and-tell presentation of a tool created here at UCSF built to bridge this gap between books and the bedside, DxTutor AI. Our presentation highlights the transformative capacity of AI simulation, aiming to inspire educators to reimagine medical education and simulation.
DxTutor AI is an interactive platform that allows students to practice and receive feedback across the entire patient encounter, from taking the history, performing the exam, ordering and interpreting labs and imaging, and creating an assessment and plan. After selecting a specialty and diQiculty, students speak to an AI patient to practice history-taking skills and choose the areas to perform a physical exam on a virtual avatar. Just like in real life, students exercise clinical judgment to order the necessary labs and imaging, rather than passively receiving pre-selected results. Students list their diQerential diagnosis and next steps and then receive personalized feedback on the entire encounter, from their bedside manner and history-taking skills, appropriateness of lab and imaging orders, to their diQerential diagnoses and next steps. It’s entirely powered by AI, from the case itself, the interactive patient, the labs and imaging, and the personalized feedback.
Our presentation will begin with a discussion on how generative AI was used to create DxTutor AI and its components. We will then perform a real-time demonstration of a complete patient encounter, followed by an interactive segment where audience members can engage with the AI patient themselves and review the AI-generated feedback. We will wrap up with a Q&A regarding this tool and AI-based simulation.
We are excited about the opportunity to present DxTutor AI at the UCSF AI and Education Symposium and advance the conversation on the future of AI in medical education.
Contact
Cat Blebea, [email protected]
Dan Qazi, [email protected]