ChatGPT in maternal-fetal medicine practice: a primer for clinicians.

ABSTRACT

ChatGPT (Generative Pre-trained Transformer), a language model that was developed by OpenAI and launched in November 2022, generates human-like responses to prompts using deep-learning technology. The integration of large language processing models into healthcare has the potential to improve the accessibility of medical information for both patients and health professionals alike. In this commentary, we demonstrated the ability of ChatGPT to produce patient information sheets. Four board-certified, maternal-fetal medicine attending physicians rated the accuracy and humanness of the information according to 2 predefined scales of accuracy and completeness. The median score for accuracy of information was rated 4.8 on a 6-point scale and the median score for completeness of information was 2.2 on a 3-point scale for the 5 patient information leaflets generated by ChatGPT. Concerns raised included the omission of clinically important information for patient counseling in some patient information leaflets and the inability to verify the source of information because ChatGPT does not provide references. ChatGPT is a powerful tool that has the potential to enhance patient care, but such a tool requires extensive validation and is perhaps best considered as an adjunct to clinical practice rather than as a tool to be used freely by the public for healthcare information.