Reply to Comment: Evaluation of Artificial Intelligence-generated Responses to Common Plastic Surgery Questions

Le résumé fourni par la source

Dear Sir, We appreciate the comments from Daungsupawong and Wiwanitkit1 about our study comparing ChatGPT and Bing’s responses to common plastic surgical queries.2 We agree that the information on the Food and Drug Administration and American Society of Plastic Surgeons websites focuses on breast implants and related disorders. In part to address this concern, the second part of our study assessed the ability of these artificial intelligence (AI) bots to answer multiple-choice questions taken from the in-service examination for plastic surgery residents. This included questions on a range of plastic surgery topics beyond breast implants. We also agree that the Food and Drug Administration and American Society of Plastic Surgeons websites may not always contain entirely current or accurate information, but selected these sources because they are reputable and evidence-based. Although the complexity and difficulty of multiple-choice questions vary, no entirely unbiased metrics are available to evaluate these aspects. Some of the questions we selected required fact-based answers, whereas others required second or third-order analysis. Not surprisingly, the AI platforms provided less reliable responses to questions involving nuanced decision-making. Studies comparing machine-generated responses to those from residents at various levels of training could inform the utility of AI technology in medical education and training. Ours was a pilot study designed to explore the ability of two commonly used AI platforms to respond to questions that both patients and clinicians typically pose and to illustrate the potential limitations of this technology. As we discussed, although queries explored using internet search engines point to information from multiple sources that users can selectively pursue, AI bots such as the two we tested typically present single summary responses as definitive answers. Further research is needed to determine the accuracy and reliability of AI in plastic surgery and other fields of healthcare. Regulatory bodies, professional organizations, practitioners, and the public must recognize both the potential applications and limitations of this emerging technology. DISCLOSURE The authors have no financial interest to declare in relation to the content of this article.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.