Abstract

Chat Generative Pre-trained Transformer (ChatGPT), an artificial intelligence chatbot, produces detailed responses and human-like coherent answers, and has been used in the clinical and academic medicine. To evaluate its accuracy in regional anesthesia topics, we produced a ChatGPT review on the addition of dexamethasone to prolong peripheral nerve blocks. A group of experts in regional anesthesia and pain medicine were invited to help shape the topic to be studied, refine the questions entered in to the ChatGPT program, vet the manuscript for accuracy, and create a commentary on the article. Although ChatGPT produced an adequate summary of the topic for a general medical or lay audience, the review that were created appeared to be inadequate for a subspecialty audience as the expert authors. Major concerns raised by the authors included the poor search methodology, poor organization/lack of flow, inaccuracies/omissions of text or references, and lack of novelty. At this time, we do not believe ChatGPT is able to replace human experts and is extremely limited in providing original, creative solutions/ideas and interpreting data for a subspecialty medical review article.