Tijana Spasojevic, Aleksandar Knezevic
Chronic pain presents a complex, multifactorial challenge in rehabilitation medicine, requiring nuanced, person-centered interventions. Artificial intelligence (AI) has emerged as a potential tool for personalizing diagnosis, predicting outcomes, and optimizing therapy, yet its integration into clinical practice remains fragmented and ethically underdeveloped. This study provides a critical, evidence-based synthesis of AI in chronic pain rehabilitation, drawing on findings from implementation studies, patient-experience research, and equity-by-design initiatives to assess where and how AI can responsibly enhance rehabilitation practices. A narrative review and critical synthesis were conducted using literature from PubMed, Scopus, and Web of Science published between 2015 and 2025. Five thematic domains emerged: spectrum of clinical maturity, patient experience and therapeutic alliance, algorithmic equity and bias, regulatory governance, and economic viability. AI applications ranged from low-touch mobile apps using sensor data to integrated Clinical Decision Support systems. Key barriers included algorithmic opacity, patient emotional burden, such as digital fatigue, biased datasets, and fragmented reimbursement models. Promising pathways included explainable and auditable AI, stakeholder co-creation, and value-based pricing. The findings emphasize that AI adoption success depends less on algorithmic sophistication and more on contextual fit, transparency, and ethical design. AI in pain rehabilitation must evolve from technological novelty to ethical necessity by aligning its development with clinical values and human well-being. Transparent systems, participatory development, dynamic oversight, and equitable access are essential for ensuring that AI enhances rather than amplifies disparities in healthcare.