Secondary Collaborators
Sarah Gustafson, Assistant Professor
Peter Quiros, Professor
Anders Garlid, Postdoctoral scholar
AI for Learning Branch: AI Clinical Reasoning Tutor
Recent widespread generative AI(Gen AI) adoption before AI guidance/regulation has posed unprecedented challenges in maintaining the quality and integrity of medical education. Thankfully, Generative AI can also accelerate and personalize medical education. We believe the current environment presents an excellent opportunity to create a Center of Excellence for AI driven innovations at the UCLA School of Medicine. To this end, we propose a dual initiative project:
Common Concerns with AI Use
1. Student Gen AI use without Guidance: The effectiveness of our traditional teaching and assessment approaches are challenged by widespread student use of generative AI on homework and learning, the lack of AI use guidance, and our inability to accurately detect student AI use. AI clinical decision support tools are being implemented more quickly than regulation or external validation can keep up, sparking concerns surrounding bias, privacy, and security. It is critical for future physicians to have advanced clinical reasoning skills and the ability to determine if, when, and how to use AI.
2. Faculty AI Discomfort: Internal faculty surveys show high levels of burnout and discomfort with Gen AI use. They are asking for guidelines and training.
Proposed Solutions:
1. AI for Learning: Our AI Clinical Reasoning (CR) tutor simulates students’ experiences in clinical rotations and provides actionable feedback on each student’s clinical reasoning skills. This will replace much of the traditional written homework, which did not emphasize CR as much, and was easy cognitively offload to Gen AI. We propose to test our AI CR tutor in a randomized control trial to compare its effectiveness to our standard curriculum.
2.AI for Teaching: An AI in Medical Education Academy + Forum for Faculty – Faculty will receive AI in Medical Education Workshops to encourage the use of Gen AI to improve teaching methods in a classroom with students of varying needs and abilities. Concurrently, faculty can model safe and effective Gen AI use to students. Faculty will also receive prebuilt generative AI chatbots that provide suggestions for written homework and examination feedback.