Enhancing cognitive engagement in fostering higher-order thinking during student discussion in medical education: a comparative study of GenAI-led, teacher-led, and blended models using Bloom's taxonomy (Preprint)

BACKGROUND Background: The integration of artificial intelligence (AI) with human expertise in blende learning models is widely recommended to address the limitations of using generative AI (GenAI) alone in medical education, especially in supporting high order thinking among peer discussions for case-based learning; however, empirical evidence supporting this approach remains limited. OBJECTIVE Objective: This study aimed to evaluate engagement and cognitive levels in medical students' discussions for thyroid cancer cases using Bloom's taxonomy under three instructional models: teacher-led, GenAI-led, and human-AI blended. METHODS Methods: In this quasi-experimental study, 120 medical students were assigned to analyze and evaluate thyroid cancer cases in the three models. RESULTS Results: Surprisingly, quantitative analysis revealed that the GenAI-led group made significantly more frequent contributions than the blended model (Δ+133%, P