Article ID: 2026.01012
Older adults living with dementia and coexisting functional disability often require care across hospitals, community services, and homes, yet information, accountability, and intervention capacity remain fragmented. This narrative review focuses on community-dwelling older adults with diagnosed dementia and limitations in basic or instrumental activities of daily living. Artificial intelligence (AI), including machine learning, natural language processing, and generative systems, may support risk interpretation and decision-making. Wearables, ambient sensors, telemedicine, and remote-care platforms provide supporting digital infrastructure but do not necessarily use AI. We synthesize evidence on these distinct components and propose CARE–AI (Capture, Analyze, Respond, and Evaluate and adapt, supported by AI), a conceptual nursing-led framework for hospital–community–home integration. Its organizational contribution is to connect algorithmic signals to named nursing review, multidisciplinary response, confirmed hand-offs, and outcome feedback while preserving personhood and caregiver partnership. Evidence predominantly concerns technical performance, feasibility, and selected clinical or psychosocial outcomes; findings from broader digital interventions cannot establish AI-specific benefit. Evidence from frailty, mild cognitive impairment, and long-term care populations is qualified for its indirect applicability. Sustained effects on function, health-care use, caregiver strain, cost, and nursing workload remain uncertain. CARE–AI as a complete framework has not been empirically evaluated. Fewer falls, fewer avoidable readmissions, and reduced caregiver strain are intended outcomes for prospective evaluation, not demonstrated effects. Evaluation should address effectiveness, equity, safety, continuity, and net nursing workload, with human oversight, explainability, consent, and appeal mechanisms built into deployment.
