Evaluation of the Accuracy of Artificial Intelligence (AI) Models in Dermatological Diagnosis and Comparison With Dermatology Specialists

Abstract

Recent advances in generative artificial intelligence (AI) have expanded its applications in diagnostic support within dermatology, but its clinical accuracy requires ongoing evaluation. This study compared the diagnostic performance of three advanced AI models, ChatGPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro, with that of board-certified dermatologists, using a dataset of 30 cases encompassing a variety of dermatological conditions. The AI models demonstrated diagnostic accuracy comparable to, and sometimes exceeding, that of the specialists, particularly in rare and complex cases. Statistical analysis revealed no significant difference in accuracy rates between the AI models and dermatologists, indicating that AI may serve as a valuable supplementary diagnostic tool in dermatological practice. Limitations include a small sample size and potential selection bias. However, these findings underscore the progress in AI’s diagnostic capabilities, supporting further validation with larger datasets and diverse clinical scenarios to confirm its practical utility.