Using ChatGPT to Generate Research Ideas in Dysphagia: A Pilot Study.

Abstract

Current research in dysphagia faces challenges due to the rapid growth of scientific literature and the interdisciplinary nature of the field. To address this, the study evaluates ChatGPT, an AI language model, as a supplementary resource to assist clinicians and researchers in generating research ideas for dysphagia, utilizing recent advancements in natural language processing and machine learning. The research ideas were generated through ChatGPT's command to explore diverse aspects of dysphagia. A web-based survey was conducted, 45 dysphagia experts were asked to rank each study on a scale of 1 to 5 according to feasibility, novelty, clinical implications, and relevance to current practice. A total of 26 experts (58%) completed the survey. The mean (± sd) rankings of research ideas were 4.03 (± 0.17) for feasibility, 3.5 (± 0.17) for potential impact on the field, 3.84 (± 0.12) for clinical relevance, and 3.08 (± 0.36) for novelty and innovation. Results of this study suggest that ChatGPT offers a promising approach to generating research ideas in dysphagia. While its current capability to generate innovative ideas appears limited, it can serve as a supplementary resource for researchers.