ABSTRACT Background and Aims ChatGPT, an AI language model developed by OpenAI, has emerged as a transformative tool in the healthcare sector. Its ability to process vast amounts of medical data and generate human‐like responses offers significant potential for enhancing patient engagement and supporting healthcare professionals. However, despite its advantages, there are significant limitations. This scoping review aims to critically evaluate both the advantages and limitations associated with integrating ChatGPT into healthcare. Methods In this scoping review, we searched for the original articles published in online databases (PubMed, Web of Science, Scopus, and Google Scholar) from January 1, 2020, to May 30, 2023 using relevant keywords (“ChatGPT,” “Health,” “Advantage,” and “Limitation”) in English language. Data collection was done using a researcher's checklist using Excel version 2019 software. Results Out of 4982 articles found, 28 articles were included in the study. All articles were conducted in 2023. The type of study in more than half of the articles was cross‐sectional. Advantages and limitations extracted from the studies were grouped into eight and nine categories, respectively. The most common advantages were “Clinical Decision Support (CDS)” and “improved medical education,” which included enhanced teaching methods and curriculum development. Specific benefits in medical education were comprehensive drug counseling, understanding of radiation oncology, and predicting drug‐drug interactions. Research assistance benefits were also notable, such as facilitating data analysis, selecting strong research ideas, and aiding in academic writing. The most frequent limitations were “knowledge limitations and accuracy” and “reliability.” Conclusion This scoping review has provided a comprehensive analysis of the advantages and limitations of ChatGPT in the healthcare sector. While the model shows significant promise in enhancing CDS and improving medical education, it is essential to remain vigilant about its limitations, particularly concerning knowledge accuracy and reliability.
Ghasemi et al. (Mon,) studied this question.