Potential of AI Chatbots in Online Hair Transplantation Consultations: A Multi-metric Assessment of Three Models.

[BACKGROUND] The use of artificial intelligence (AI) chatbots has demonstrated considerable promise in assisting medical consultations. However, their potential for application in online hair transplantation consultations remains largely unexplored.

[OBJECTIVES] This study aims to assess the effectiveness of AI chatbots in responding to patient inquiries during online hair transplantation consultations.

[METHODS] We evaluated responses to 10 common patient questions collected from online hair transplantation clinics, comparing answers generated by three AI chatbots-ChatGPT-4o mini, Claude 3.5 Sonnet, and Gemini Advanced-with those from senior surgeons. Each response was scored based on medical accuracy, empathy, understandability, actionability, and readability, with a focus on determining how well AI can match or exceed human expert performance.

[RESULTS] All three AI chatbots matched or outperformed the response capabilities of senior surgeons in medical accuracy, empathy, understandability, actionability, and readability. Among them, Gemini Advanced showed the most comprehensive advantages, including significantly higher scores in medical accuracy (4.5 vs. 3.9, P < .001), empathy (4.9 vs. 2.5, P < .001), and understandability (82.7% vs. 63.9%, P < .001). Additionally, Gemini Advanced demonstrated a lower Flesch-Kincaid Grade Level (10.5 vs. 18.7, P < .001) and higher Flesch Reading Ease Score (40.3 vs. 16.2, P < .001), suggesting better readability.

[CONCLUSION] AI chatbots show strong potential for use in online hair transplantation consultations, providing accurate, empathetic, and easily understandable responses. Nevertheless, challenges such as privacy concerns, ethical considerations, and potential biases need to be addressed before their adoption in clinical practice.

[LEVEL OF EVIDENCE IV] This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .