Scoping Review Maps What AI Caregiving Tools Actually Do, and How Thin the Evidence Still Is

A new scoping review catalogs 43 studies of AI tools built for family caregivers of people with Alzheimer's disease and related dementias, from robots to chatbots. The authors found a wide range of functions on offer and called for rigorous, long-term testing that so far has not been done.

A scoping review published online July 17, 2026, in the Journal of Gerontological Nursing identified 43 studies of artificial intelligence technologies built to support family caregivers of people with Alzheimer's disease and related dementias (ADRD), and concluded that future research needs to move toward rigorous, long-term empirical evaluation of these tools. The review, by Alison H. Davis of Auburn University College of Nursing and Mary Dioise Ramos of the Louisiana State University Health Sciences Center-New Orleans School of Nursing, searched the CINAHL and PubMed databases and synthesized study designs, the AI technologies used, caregiver populations, care settings, outcomes measured, and key findings across the 43 studies.

The technologies identified fell into three broad categories: physical robots, conversational agents such as chatbots, and predictive algorithms that flag changes in a person's condition. Across the studies, these tools were used to support caregivers in four ways: education about the disease and caregiving tasks, direct task assistance, social support, and clinical management such as monitoring symptoms or logging behavior changes. The review's own framing, a scoping review, means its purpose was to map what exists and where the gaps are, not to test whether any single tool works.

The authors' conclusion is measured rather than promotional: as the population of older adults with ADRD grows, so will demand for caregiver support, and current research already shows a wide range of AI applications in development or use. But the same conclusion states plainly that future work should focus on rigorous, long-term empirical evaluation, a signal that the evidence needed to say these tools change outcomes for caregivers is not yet in hand. That gap is not new. A 2020 systematic review of 30 earlier studies on AI for ADRD caregivers, published in the Journal of Medical Internet Research, found that most studies were qualitative or small cross-sectional designs that reported favorable impressions but carried a low level of evidence overall.

Neither review, the new one or the one from 2020, reports that any AI caregiving tool has been shown in a large, controlled trial to reduce caregiver burden, delay a move to institutional care, or improve a person's symptoms. Families weighing an app, a sensor, or a companion robot are, for now, choosing among products that have mostly been described and categorized, not proven.