Assessment of exercise accuracy and explanation with artificial intelligence: A study on cervical exercises for nonspecific neck pain

Authors:
Publisher website
Google Scholar
Abstract
Aim: This study evaluated the accuracy and explanation of ChatGPT 4.0 in providing cervical exercise instructions to individuals with nonspecific neck pain. Materials and Methods: The study was conducted between October 2024 and January 2025. Frequently recommended cervical exercises were selected through a literature review. ChatGPT 4.0 was asked to explain these exercises in different positions. Accuracy and explanation quality were assessed using the Learning Objective Review Tool by an expert team. Statistical analysis was performed using JASP software to calculate means and standard deviations for scores. Results: Craniocervical flexion exercises in the supine position had the highest accuracy score (3.65±0.518). However, explanation scores were lower, especially for exercises in standing and quadruped positions (1.5±0.535). Experts suggested corrections for clarity and detail. Conclusion: ChatGPT 4.0 shows promise for exercise recommendations, but accuracy and explanation need improvement. Future research should focus on integration of artificial intelligence collaboration with developers and healthcare professionals.
Funding Information
  • This research did not receive financial support from any institution or individual.

This publication has 23 references indexed in Scilit: