The design, practice, and evaluation of AI-assisted cloud platform in medical biochemistry education

Abstract Aim This study evaluated the effectiveness of a multi-module AI-assisted teaching cloud platform in medical biochemistry education. Background Biochemistry is a core medical course characterized by abstract, knowledge-dense content that challenges traditional teaching. Empirical evidence on systematically integrating multiple AI interaction modules within a complete biochemistry course remains limited. Methods A 20-week controlled experiment was conducted with 151 eight-year clinical medicine students, assigned to an AI platform group (n = 69, using the Chaoxing AI platform with Basic Learning, Case Discussion, Virtual Practice, and Comprehensive Assessment modules) or a control group (n = 82, using traditional tools). Final exam scores, higher-order thinking question performance, delayed post-test (6 weeks post-course) for knowledge retention, platform usage data, and questionnaire responses were collected and analyzed. Results The AI group significantly outperformed the control group on higher-order thinking questions (7.52 ± 0.78 vs. 5.95 ± 0.99, p = 0.015). Platform usage time positively correlated with final scores (R = 0.4082, p < 0.0001), but no significant group difference emerged in knowledge retention (65.75 ± 13.05 vs. 63.24 ± 13.42). Although 77.78% of teachers agreed the platform enriched teaching, 55.56% disagreed it reduced their workload. Conclusions The AI platform enhances higher-order thinking but does not boost basic knowledge retention or reduce teacher burden. Future optimization should target personalized algorithms and teacher training to improve human-computer collaboration.