Evaluation and Comparison of the Knowledge Levels of Current Artificial Intelligence Programs on Retinal/Vitreous Diseases and Treatment Methods

Abstract

Purpose: To evaluate the answers to multiple-choice questions about retina and vitreous diseases and treatment modalities of Chat Generative Pre-Trained Transformer (ChatGPT), Bard, and Bing artificial intelligence chatbots, examining the level of knowledge about these subjects, and investigating the existence of their superiority over each other. Methods: Forty-six questions related to retinal and vitreous diseases and treatment modalities were asked to ChatGPT, Bing, and Bard chatbots. Results: The Bing artificial intelligence chatbot correctly answered 76.1% of the questions. ChatGPT and Bard artificial intelligence chatbots correctly answered 60.9% of the questions. No statistically significant difference was observed between the rates of correct and incorrect answers to the questions on the three artificial intelligence chatbots (P = 0.206). Conclusions: Artificial intelligence chatbots can be used to access accurate information about retinal and vitreous diseases and treatment modalities. However, the information obtained may not always be correct, and care should be taken about its use and results.