Objective: This narrative review aims to evaluate the clinical effectiveness, safety, and feasibility of artificial intelligence (AI)–supported physiotherapy compared to conventional approaches. Methods: A systematic search of PubMed (2018-2025) identified comparative studies involving AI-based physiotherapy interventions. Inclusion criteria comprised original research with intervention/control groups and statistical comparisons. Methodological quality was assessed using the PEDro scale. Results: Four studies met eligibility, spanning chronic pain, post-surgical, and stroke rehabilitation contexts. AI-assisted interventions showed comparable or superior outcomes in pain reduction, functional recovery, adherence, and patient satisfaction, with no reported adverse events. Mean PEDro score was 6/10. Conclusion: AI-assisted physiotherapy appears effective, feasible, and well-accepted across diverse clinical contexts. While some interventions outperformed traditional care, others showed equivalent outcomes with improved engagement. These findings support the integration of AI into rehabilitation and highlight areas for further research.
Artiles-Sánchez et al. (Sat,) studied this question.