Abstract
Communicating pain is challenging in healthcare due to its subjective nature and patients' difficulties in describing it.
This study investigates using AI models, DALL-E and ChatGPT, to generate and interpret visual representations of different pain types-burning, radiating, shooting, stabbing, throbbing, dull, and pins and needles.
DALL-E created pictograms from specific prompts, and ChatGPT identified the pain types, achieving 6 out of 7 correct identifications.
The model's explanations highlighted its ability to recognize visual cues like shapes and colors.
A misidentification of ”shooting pain” as ”radiating pain” prompted a refinement of the image, leading to correct identification, underscoring the need for specificity in visual representations.
These findings suggest AI-generated images can enhance pain communication between patients and healthcare providers, with future research needed to optimize and validate these tools in clinical settings.
