ChatGPT Responses to Questions About Clinical Decision Making in Speech-Language Pathology

What is it about?

This paper looks at the usefulness of ChatGPT, an artificial intelligence software, as a clinical tool for speech-language pathologists. To see how accurate ChatGPT is at answering clinical questions, nine university-based communication sciences and disorders (CSD) researchers with expertise in different areas developed questions of varying difficulty—basic, advanced, research-based clinical questions, and general CSD. They asked ChatGPT these questions and then rated its answers for accuracy on a 6-point scale and for completeness on a 3-point scale. The results show that, on average, ChatGPT can provide highly accurate and complete answers to many questions (based on expert ratings). However, the spread of scores also shows that more than half of its answers were not entirely accurate, and nearly half were not fully complete.

Featured Image

Photo by Levart_Photographer on Unsplash

Photo by Levart_Photographer on Unsplash

Why is it important?

This study shows that although ChatGPT can quickly provide answers that experts find accurate, it cannot be relied on for fully accurate or complete information, especially for complex clinical questions. Using ChatGPT to make clinical decisions can be risky for students or less experienced clinicians who may not recognize mistakes in its responses.

Perspectives