Thu, Sep 3, 2026
Volume 16 - Special Issue(Exercise Therapy in Action)-In press
PTJ 2026, 16 - Special Issue(Exercise Therapy in Action)-In press: 0-0 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
rajabi R. Large Language Models in Corrective Exercise Prescription: Current Evidence, Clinical Challenges, and Future Directions. PTJ 2026; 16 (S1)
URL: http://ptj.uswr.ac.ir/article-1-957-en.html
URL: http://ptj.uswr.ac.ir/article-1-957-en.html
Abstract: (3 Views)
Corrective exercise is an assessment-driven approach designed to address movement dysfunctions, postural deviations, neuromuscular impairments, and related functional limitations through individualized exercise selection, dosage, progression, and reassessment. The emergence of large language models (LLMs), particularly ChatGPT, has created new opportunities for exercise prescription and rehabilitation by enabling rapid synthesis of scientific information, generation of structured exercise programs, patient education, and clinical documentation. Current evidence suggests that LLMs can generate generally coherent exercise recommendations incorporating variables such as frequency, intensity, time, type, volume and progression (FITT-VP); however, their accuracy, individualization, and clinical reliability remain variable.
Applying LLMs to corrective exercise presents additional challenges because effective intervention depends on accurate static and dynamic posture assessment, interpretation of clinically meaningful impairments, individualized exercise selection, and continuous adaptation. Text-based LLMs cannot independently perform physical examinations, directly assess movement quality, or reliably determine the clinical significance of observed postural characteristics. Consequently, a plausible AI-generated exercise program should not be considered equivalent to a clinically validated intervention.
Recent advances in multimodal artificial intelligence provide complementary capabilities. Computer vision and human pose estimation can identify anatomical landmarks and estimate joint angles, segment relationships, and movement kinematics, while markerless cameras, depth sensors, and wearable inertial measurement units (IMU) can facilitate objective movement assessment, exercise monitoring, and real-time feedback. These technologies may provide the movement data required to enhance LLM-assisted decision support.
This mini-review examines current evidence, clinical challenges, and emerging opportunities for AI-assisted corrective exercise prescription. A human-AI collaborative model is proposed in which multimodal AI provides objective movement assessment and monitoring, LLMs support evidence synthesis and candidate exercise-program generation, and qualified professionals retain responsibility for clinical interpretation, prescription, progression, and safety. Future research should prioritize standardized evaluation, multimodal integration, and clinical outcome validation.
Applying LLMs to corrective exercise presents additional challenges because effective intervention depends on accurate static and dynamic posture assessment, interpretation of clinically meaningful impairments, individualized exercise selection, and continuous adaptation. Text-based LLMs cannot independently perform physical examinations, directly assess movement quality, or reliably determine the clinical significance of observed postural characteristics. Consequently, a plausible AI-generated exercise program should not be considered equivalent to a clinically validated intervention.
Recent advances in multimodal artificial intelligence provide complementary capabilities. Computer vision and human pose estimation can identify anatomical landmarks and estimate joint angles, segment relationships, and movement kinematics, while markerless cameras, depth sensors, and wearable inertial measurement units (IMU) can facilitate objective movement assessment, exercise monitoring, and real-time feedback. These technologies may provide the movement data required to enhance LLM-assisted decision support.
This mini-review examines current evidence, clinical challenges, and emerging opportunities for AI-assisted corrective exercise prescription. A human-AI collaborative model is proposed in which multimodal AI provides objective movement assessment and monitoring, LLMs support evidence synthesis and candidate exercise-program generation, and qualified professionals retain responsibility for clinical interpretation, prescription, progression, and safety. Future research should prioritize standardized evaluation, multimodal integration, and clinical outcome validation.
Keywords: Large language models, ChatGPT, Corrective exercise, Exercise prescription, Artificial intelligence, Clinical decision support, Rehabilitation
Type of Study: Editorial |
Subject:
Sport injury and corrective exercises
Received: 2026/08/11 | Accepted: 2026/08/23 | Published: 2026/06/22
Received: 2026/08/11 | Accepted: 2026/08/23 | Published: 2026/06/22
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 (CC-By-NC 4.0), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
