Exploring the competence of artificial intelligence programs in the field of oculofacial plastic and orbital surgery

Abstract

Aims: It aims to evaluate the knowledge level of ChatGPT, Bing, and Bard artificial intelligence chatbots developed based on Large Language Models (LLM) about oculofacial plastic surgery and to investigate the presence of superiority over each other. Methods: Twenty-nine questions that tested knowledge about oculofacial plastic and orbital surgery were taken from the study questions section of the American Academy and Ophthalmology 2022-2023 Basic and Clinical Science Course Oculofacial Plastic and Orbital Surgery. The questions were asked to ChatGPT, Bing, and Bard programs, which are current artificial intelligence chatbots. The questions were classified as either correct or incorrect. Results: ChatGPT gave 44.8% correct answers, Bing 48.3% correct answers, and Bard 58.6% correct answers to 29 questions about artificial intelligence chatbots. No statistical difference was observed between the rates of correct and incorrect answers given by 3 the intelligence programs (p=0.609, Pearson’s chi-squared test). Conclusion: The use of artificial intelligence to access information regarding oculofacial plastic and orbital surgery may provide limited benefits. Care should be taken in terms of accuracy and timeliness when evaluating the results of artificial intelligence programs.