Evaluation of the Impact of AI-Driven Personalized Learning Platform on Medical Students' Learning Performance

Yajun Chen^*

• Heilongjiang Nursing College, HarBin City, China

Objective:This study aims to evaluate the comprehensive impact of an AI-driven personalized learning platform based on the Coze system on medical students' academic performance, learning satisfaction, and self-directed learning behaviors. The goal is to explore its practical value in medical education and provide evidence for educational digital transformation. Methods: A 12-week prospective randomized controlled trial was conducted with 40 medical undergraduates, stratified by baseline academic levels and randomly assigned to an experimental group (n=20, AI-based Coze learning) and a control group (n=20, traditional teaching). The intervention included: (1) Dynamic learning path optimization based on weekly diagnostic tests; (2) Real-time emotion detection using natural language processing with adaptive motivational feedback; (3) BERT-based intelligent resource recommendation from a 2,800-case database; (4) Embedded virtual case simulations offering real-time clinical guidance. The control group received conventional lectures (4 hours/week) with standard materials. Data collected included academic scores (3 standardized tests, Cronbach's α=0.89), satisfaction (5-point Likert scale, α=0.84), and behavioral indicators (learning duration, class interaction frequency, literature reading volume). Analyses were conducted via SPSS 26.0.The experimental group outperformed controls in test scores (84.47 vs. 81.72, p=0.034, d=0.72), satisfaction (17.45 vs. 16.05, p=0.042, d=0.36), and classroom participation (16.05 vs. 7.40 times/session, p=0.026, d=0.83). Self-directed learning improved with a 41.5% increase in daily study time and a 48.3% rise in literature reading (p=0.008, d=1.14). Both metrics positively correlated with academic performance.The AI-powered Coze platform significantly enhances learning outcomes, satisfaction, and self-directed behaviors through adaptive content delivery, affective support, and data-driven insights. It holds substantial promise for intelligent medical education reform, though long-term effects and ethical considerations merit further exploration.