A comparative evaluation of ChatGPT-assisted and traditional methods of teaching diagnostic assessments to medical students

The rapid development of artificial intelligence (AI) technology has brought new opportunities and challenges to the ever-evolving field of medical education. However, the specific impact of AI on the teaching of diagnostic assessments to medical students requires further investigation. The purpose of this study was to evaluate the effectiveness of ChatGPT-assisted teaching of diagnoses. Eighty medical students were randomly divided into 2 groups; 1 group received ChatGPT-assisted training, and the other group received conventional instruction. We evaluated the effectiveness of both teaching methods using a theoretical examination and the Mini-Clinical Evaluation Exercise (Mini-CEX). We also assessed satisfaction with the ChatGPT-assisted teaching method using a structured survey. The ChatGPT-assisted students scored significantly higher on the theoretical knowledge examination compared to the traditional teaching group ( t = 6.34, P < .001). The ChatGPT-assisted group showed significant improvement on the Mini-CEX examination, particularly on medical history taking (Χ 2 = 5.115, P = .024), doctor-patient communication (Χ 2 = 5.051, P = .025), professionalism (Χ 2 = 4.073, P = .044), clinical judgment (Χ 2 = 4.503, P = .034), and overall capabilities (Χ 2 = 7.218, P = .007). The majority of participants rated the ChatGPT teaching method highly and provided positive feedback, highlighting the benefits of interactive learning and skill acquisition. ChatGPT-assisted teaching not only improved the students’ scores on the theoretical knowledge of diagnostic assessments, but also enhanced the clinical knowledge of the medical students. The positive feedback of the ChatGPT-assisted group shows that this method has the potential to supplement and enhance conventional teaching methods in medical education. These results support further expansion and the application of AI in medical education.