Using Artificial Intelligence in Clerkship Learner Assessment: A CERA Study

Background and Objectives: The integration of artificial intelligence (AI) in medical education primarily has focused on clinical applications, with limited investigation into its role in learner assessment. This study explores AI usage among family medicine clerkship directors and their perspectives on AI’s potential in competency-based medical education assessment. Methods: Data were collected through the 2024 Council of Academic Family Medicine Educational Research Alliance survey of family medicine clerkship directors. The survey was distributed from June four to July 12, 2024, to 173 directors, achieving a 52.6% (91/173) response rate. We used multivariable linear regression to analyze the relationship between AI usage and a composite favorability score for AI integration in student assessment, adjusting for covariates such as clerkship design and director tenure. Results: All respondents were physicians leading mandatory clerkships. Female and underrepresented minority directors were more likely to report no prior AI use. Multivariable analysis demonstrated a significant positive association between AI usage and favorability toward AI in learner assessment (coefficient: 2.601; 95% CI: 1.246–3.956; P<.001), even after adjusting for multiple comparisons. Feedback quality, organizational support, and policies did not significantly impact favorability. Conclusions: AI exposure was significantly associated with favorable attitudes toward AI in learner assessment, while organizational factors had no significant effect. Future studies with larger samples and longitudinal designs may clarify how institutional support and increasing AI exposure influence attitudes over time, informing best practices for AI adoption in medical education.