Related Experiment Videos
Artificial Intelligence in Exercise Programming and Coaching: Opportunities and Limitations
Susannah L Reiner1,2, Regis C Pearson3, Rachelle A Reed4
1Department of Interprofessional Health Sciences and Health Administration, Seton Hall University, South Orange, NJ.
Current Sports Medicine Reports
|August 5, 2026
Summary
Artificial intelligence (AI) shows promise for exercise prescription, but real-world use lags behind research. More validation is needed for AI tools to effectively augment coaching and improve long-term outcomes.
Area of Science:
- Exercise Science
- Artificial Intelligence
- Sports Technology
Background:
- Artificial intelligence (AI) offers transformative potential for exercise prescription and coaching via machine learning, deep learning, and large language models.
- A significant gap exists between peer-reviewed evidence and the deployment of AI in commercial fitness platforms.
- Recent literature (2023-2025) highlights the growing integration of AI in exercise programming.
Purpose of the Study:
- To review and assess recent literature on data-driven programming approaches for exercise prescription.
- To identify current capabilities and limitations of AI in exercise science.
- To provide recommendations for future research and development.
Main Methods:
- A narrative review of peer-reviewed literature published between 2023 and 2025.
- Literature search focused on four key themes: wearable data integration, predictive modeling, AI coaching, and large language models.
- Synthesis of findings regarding the application and effectiveness of AI in exercise prescription.
Main Results:
- AI demonstrates feasibility in activity recognition, workload estimation, and short-term performance prediction using wearable data.
- Chatbots and virtual coaches show potential for user engagement and guidance.
- Significant gaps persist in evaluating closed-loop adaptive programming, long-term effectiveness, and behavioral outcomes.
- Commercial platforms often integrate multimodal data and evolve rapidly, frequently lacking formal validation.
Conclusions:
- AI is a feasible tool for specific aspects of exercise prescription, such as data analysis and short-term predictions.
- Translational research models are crucial to bridge the gap between academic findings and industry application.
- Future advancements should prioritize transparency, human-in-the-loop systems, and rigorous evaluation of AI's long-term impact on health and performance.
- AI should be viewed as a complementary tool to enhance, not replace, professional human coaching.
