Understanding ChatGPT for evidence-based utilization in interventional radiology

• Published in: Clinical imaging 2024-04, Vol.108, p.110098-110098, Article 110098
• Main Authors: Campbell, Warren A., Chick, Jeffrey F.B., Shin, David, Makary, Mina S.
• Format: Article
• Language: English
• Subjects: ChatGPT
• ISSN: 0899-7071; 1873-4499
• DOI: 10.1016/j.clinimag.2024.110098

Advancement in artificial intelligence (AI) has the potential to improve the efficiency and accuracy of medical care. New techniques used in machine learning have enhanced the functionality of software to perform advanced tasks with human-like capabilities. ChatGPT is the most utilized large language model and provides a diverse range of communication tasks. Interventional Radiology (IR) may benefit from the implementation of ChatGPT for specific tasks. This review summarizes the design principles of ChatGPT relevant to healthcare and highlights activities with the greatest potential for ChatGPT utilization in the practice of IR. These tasks involve patient-directed and physician-directed communications to convey medical information efficiently and act as a medical decision support tool. ChatGPT exemplifies the evolving landscape of new AI tools for advancing patient care and how physicians and patients may benefit with strategic execution. •ChatGPT is a language model trained to generate text to perform complex tasks.•ChatGPT implements a transformer learning model to generate bodies of text.•ChatGPT integration has the potential to automate or assist vital clinical and medical functions in interventional radiology.•Preliminary evidence suggests ChatGPT or chatbot technology can improve communication between patients and physicians.•Future research is required to evaluate patient safety and clinical efficacy of each ChatGPT iteration prior to implementation.