Artificial Intelligence-Assisted Fall Prevention in Older Adults - Risk Identification, Individualized Intervention, and Clinical Translation

Abstract

Falls are a leading cause of disability, hospitalization, and mortality in later life. Although traditional fall-prevention assessment already includes relatively mature scales and physical examination, it is still largely based on one-time screening and may miss fluctuating risks in home and community settings. By reviewing published work, this paper discusses how AI may shift traditional fall prevention from post-event recognition to pre-event prediction and what applications this may enable in VR and AR training, robot rehabilitation, and closed-loop management across home, community, and hospital settings. Current clinical use still requires stronger external validation, better explainability, clearer privacy and bias safeguards, more persuasive cost-effectiveness evidence, and smoother workflow integration. AI has real potential for objective assessment and dynamic management, but its current clinical soundness remains uncertain.