AI-Driven Virtual Patient Simulations for Communication Training in Medical Students: A Pilot Study

Background Effective communication is essential in medical practice, yet traditional training methods face limitations, including resource constraints, variability in instructional quality, and inconsistent feedback. Artificial Intelligence (AI)-driven tools provide scalable and interactive solutions for developing communication skills in a structured manner. This pilot study evaluated the feasibility and effectiveness of an AI-based virtual patient simulation tool for enhancing communication skills in medical students. Specifically, the study examined whether the tool would demonstrate high usability among participants and whether the AI-generated feedback would be perceived as clear, well-structured, and useful for the development of communication competencies. Methods A total of 102 medical students from the University of Málaga participated in a GPT-4-powered virtual patients’ interactions across two standardized clinical cases. Usability, satisfaction, and feedback quality were assessed via post-session surveys, Likert-scale ratings, and qualitative feedback. Statistical analyses included a binary logistic regression analysis that was conducted to examine whether gender influenced the perceived need for prior training before using the tool. Results Binary logistic regression revealed that gender was a significant predictor of perceived need for prior training. Women were significantly less likely than men to consider such training necessary (β = − 1.24, OR = 0.29, SE = 0.47, z = − 2.63, p = 0.008), indicating a 71% reduction in odds. Most participants (60.8%) reported no need for prior training, and 86% found the tool easy or too easy indicating high intuitive usability Conclusions Findings confirm the feasibility, high usability, and effectiveness of AI-driven virtual patient simulations, with significant differences in perceived feedback quality across dimensions. Further research should explore longitudinal skill retention, hybrid models integrating human-led debriefing, and AI-enhanced nonverbal communication analysis to refine AI-driven communication training, while addressing ethical concerns such as bias, misinformation, and the limitations of AI in providing emotionally contextualized feedback