Artificial Intelligence and the Future of Computer-Assisted Medical Research and Writing

In this issue of Ophthalmic Plastic and Reconstructive Surgery, Dr. Mohammad Javed Ali’s article entitled “ChatGPT and Lacrimal Drainage Disorders: Performance and Scope of Improvement” examines the performance of the large language model (LLM), ChatGPT (generative pre-trained transformer), in the context of some lacrimal drainage disorders. The author evaluates the artificial intelligence (AI) responses to 21 prompts, such as the history of dacryocystorhinostomy surgery, lacrimal pump anatomy, silicone intubation, and the use of mitomycin C in DCR surgery. Each prompt was repeated three times to assess consistency or differences in response, and each response was evaluated by three independent lacrimal surgeons on a three-point scale. Although the model provided a generic disclaimer that it cannot give medical advice or generate professional opinions, it did go on to answer the questions in some detail. The author found that some factual inaccuracies were noted in 60% of the ChatGPT responses and several on controversial topics were generic and not evidence-based. Even so, ChatGPT was able to acknowledge its mistakes and correct itself when presented with counterarguments. While Dr. Ali recognizes that the potential of this AI chatbot to influence medicine is enormous, he concludes that “Healthcare, as it appears, is not ready for it at present…,” and there remains a need for it to be specifically trained and retrained for individual medical subspecialties.