Artificial intelligence as a tool for providing feedback in an internal medicine residency program: a survey of residents’ perceptions

Abstract

This blinded, cross-sectional survey-based study, performed at Loma Linda University Internal Medicine Residency Program, evaluated resident perceptions of educator-generated versus AI-enhanced feedback. Fifty-six resident evaluations were distributed, and 50 residents (89%) completed the survey. AI-enhanced feedback, developed using a standardized ChatGPT-4 prompt, was generally perceived as more structured, detailed, and actionable, particularly in supporting clearer improvement plans, while educator feedback was valued for its personalized nature. AI-enhanced feedback received numerically higher ratings across domains, although differences were not statistically significant in the overall cohort. These findings suggest that AI-enhanced feedback may serve as a complementary tool to enhance the perceived educational value of trainee evaluations in residency training programs.