Using Generative AI to support clinical decision making - "It ain't what you prompt it's the way that you prompt it"

This paper was penned to showcase a method of exposing physiotherapy students to profession specific opportunities and limitations associated with the use of Generative AI. Transformational teaching methods were employed.

This paper outlines a potential method of teaching about generative AI use in clinical situations to physiotherapy students through a carefully chosen case study which included covert red flags. When asked about potential diagnosis, the Gen AI output differed significantly with very minor alterations in prompts. Such changes exposed the fact that the AI platform did not understand the professional remit of a Physiotherapist and a generated a productive in class debate.

Results expose a narrow view of the physical therapy profession by AI, and the importance of carefully considering the words used in commands given to generative AI platforms.

Physiotherapists using AI in their clinical practice to aid with clinical decision making or differential diagnosis should be aware of the importance of correct prompting and the risks and limitations associated with AI use without adequate prompting skills.

With the increased use of AI in clinical practice, teaching Physiotherapists how to leverage the opportunities associated with such tools and also exposing their limitations should be considered in the context of continued professional development and Physiotherapy curriculums.