Two research teams exploring the impact of artificial intelligence (AI) in medical education assessment have been awarded the 2026 Stemmler Grants. Their projects will examine how AI can enhance formative assessment and feedback as well as support validity and reliability for emerging evaluation methods.
Established in 1995, the Edward J. Stemmler MD Medical Education Research Grants support research and development of innovative assessment approaches with the potential to advance assessment in medical education. Each year, as many as three awards of up to $150,000 are given to research teams whose work contributes to meaningful progress across the medical education continuum. By fostering this kind of research, the program reflects NBME’s commitment to improving health care outcomes and advancing the standards of the profession.
The first 2026 project, “Artificial Intelligence-Assisted Real-Time Feedback Analysis, Collection and Tracking System (FACTS),” is led by Daniel Salcedo, MD, associate professor of General Medical Sciences, and Xiaomei Song, PhD, associate professor and director of student assessment, from Case Western Reserve University School of Medicine. Their project will develop and evaluate an AI-powered platform designed to transform verbal feedback from simulation and clinical encounters into structured, actionable data. By integrating automated speech recognition, speaker diarization and natural language processing, FACTS aims to help learners and educators track performance over time, support faculty development and improve the consistency and usefulness of formative feedback.
The second project, “Validating the Virtual Rater: A Psychometric Validation of AI-Based Video Assessment Across Diverse Rubric Types,” is led by Laurah Turner, PhD, associate dean for Artificial Intelligence and Educational Informatics at the University of Cincinnati College of Medicine. This project will validate an Agent Based Video Assessment tool across admissions interviews and Objective Structured Clinical Examination recordings to better understand whether AI can serve as a reliable assessment tool across different rubric types. The study will compare AI-generated ratings with human ratings, evaluate performance across demographic subgroups, and develop evidence-based guidelines to support the valid and reliable implementation of AI in medical assessment.
As part of ongoing efforts to highlight the impact of Stemmler-supported research, NBME will continue the Stemmler Speaker Series this fall. The next virtual event will be held on Sept.29 and will feature Dr. Verity Schaye, who will discuss her Stemmler-supported work and its contributions to medical education assessment. Sign up to attend on Zoom.
Learn more about the 2026 Stemmler grantees and other past projects on the Stemmler Grants page.
