Artificial Intelligence in Physiotherapy and the Limits of Clinical Reasoning: A Narrative Review

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Objective: This narrative review examines current applications of artificial intelligence (AI) in physiotherapy and rehabilitation and evaluates the limits of AI-supported clinical reasoning, with emphasis on therapeutic touch, human interaction, ethical and medico-legal responsibility, professional autonomy, and human oversight.Materials and Methods: Literature on AI in physiotherapy and rehabilitation was reviewed using a narrative approach. Original studies, qualitative research, reviews, and conceptual publications addressing clinical decision support, movement analysis, rehabilitation monitoring, education, therapeutic touch, ethical and medico-legal issues, and human oversight were evaluated and synthesised thematically.Results: AI may support physiotherapy through clinical decision support, movement analysis, functional outcome prediction, rehabilitation monitoring, personalised treatment planning, and education. However, strong performance in diagnostic classification or structured tasks does not necessarily translate into appropriate, individualised rehabilitation management. Clinical reasoning also requires consideration of patient preferences, psychosocial factors, therapeutic touch, non-verbal communication, and changing patient responses. Concerns include data privacy, algorithmic bias, explainability, automation bias, professional autonomy, and legal responsibility.Conclusion: AI can support clinical processes but cannot replace therapeutic touch, person-centred care, clinical reasoning, or professional responsibility. It should be used as a supportive technology under clinician leadership and meaningful human oversight, rather than as an independent substitute for the physiotherapist.