Here is the clean HTML — copy from the line starting `
By Stuart Kerr` to the end:
—
By Stuart Kerr, Technology Correspondent, LiveAIWire
By 2025, one in ten older adults had access to telehealth, according to the Oxford Academic Innovation in Aging review — which also identified this figure as highlighting a significant market opportunity. The AI-powered solutions segment within elderly care is projected to reach 2.25 billion dollars by 2030, growing from 1.78 billion dollars in 2024 at a compound annual growth rate of 9.73 percent. The global elder care assistive robots market is growing at 14.8 percent annually through 2030. These numbers reflect a genuine and urgent need: an ageing global population with growing care requirements that existing healthcare systems cannot meet through human caregiving alone, against a backdrop of caregiver shortages that are accelerating in most high-income countries.
The promise of AI in elderly care is real in its outlines. Technologies that enable older adults to live independently longer — monitoring for falls, managing medication adherence, providing cognitive stimulation, alerting caregivers or emergency services when needed — address a challenge that has not changed: maintaining quality of life and safety for people whose physical and cognitive capacities are changing, often faster than their social environments can adapt. What the marketing for these technologies frequently omits is a careful account of where the evidence for specific AI applications is robust, where it is preliminary, and where the implementation risks are significant enough to require active management.
Where the Evidence Is Strongest
The clinical evidence base for AI in elderly care is most developed for fall detection, medication management, and health monitoring applications — uses where the AI is supporting or replacing well-defined clinical functions with measurable outcomes. Fall-related injuries are the leading cause of injury death in adults over 65, and the consequences of undetected falls — the older adult lying on the floor for hours or days without help — are severe and preventable. AI systems using wrist-worn accelerometers, floor-based pressure sensors, and computer vision can detect falls with accuracy above 90 percent in controlled settings, significantly exceeding what manual monitoring could achieve in home environments where falls most commonly occur.
Medication non-adherence is responsible for approximately 10 percent of hospital admissions in older adults and costs the United States healthcare system an estimated 300 billion dollars annually. AI-powered medication management systems that dispense medications at the correct time, monitor consumption, and alert caregivers to missed doses address a well-defined and measurable problem with a technology solution that is straightforward to evaluate clinically. The evidence for specific AI medication management systems that have been clinically validated is positive, though generalisation from the best-performing systems to the broader market of medication AI products is not warranted without system-specific evaluation.
Remote health monitoring through wearable devices and AI analysis is the category growing most rapidly, with the most diverse evidence base. AI systems that monitor heart rate variability, sleep patterns, activity levels, and physiological markers can identify patterns preceding health deterioration — hospitalisation-predictive signals that appear in wearable data 72 hours or more before clinical presentation. Changes in movement patterns, feeding behaviour, and daily activity can indicate cognitive decline, infection, or medication side effects before they become clinically visible. This predictive capability, applied to the right population with appropriate clinical integration, has genuine potential to reduce hospitalisation rates and enable earlier intervention.
The AI Companion Question
AI companion robots and conversational AI tools for elderly adults are the most marketed and the most contested category in AI elderly care. Loneliness among older adults is a genuine and severe public health crisis: older adults experiencing loneliness are 45 percent more likely to die prematurely, according to University of California San Francisco research. Loneliness is compared to smoking 15 cigarettes a day in terms of its impact on health outcomes. An AI companion that provides consistent conversation, cognitive stimulation, and emotional engagement for an isolated older adult has a plausible mechanism for addressing a real harm.
The clinical evidence on AI companions for elderly adults is more limited than the enthusiasm in the market suggests. Tools like ElliQ and voice-based AI interfaces have been pilot-tested among isolated seniors with some success in reducing self-reported loneliness and increasing engagement, according to the Oxford Academic review. But long-term efficacy data remains limited. Systematic reviews of AI applications to reduce loneliness in older adults found that while AI applications showed promise in reducing loneliness in the short term, the evidence base was insufficient to draw strong conclusions about sustained benefit or to establish which populations benefit most from which types of AI interaction.
The ScienceDirect study on AI companions and subjective wellbeing found that companionship-oriented chatbot use was consistently associated with lower wellbeing among users with low levels of friend-based social support — the population of isolated elderly adults that AI companions are most explicitly designed to help. The finding does not establish that AI companions harm isolated elderly adults. It establishes that the relationship between AI companionship and wellbeing is moderated by factors that are not yet well understood, and that the vulnerable population most targeted by AI companion marketing may be the population for whom the effects are most complex.
The Risks That Are Not Being Adequately Discussed
The risks of AI elderly care that marketing materials omit fall into several categories. Data privacy is the most consistently identified risk by researchers: AI elderly care systems collect continuous, intimate data about older adults’ health, behaviour, activities, and communications in their homes. The terms under which this data is stored, shared with family members, shared with healthcare providers, used to train AI models, and potentially disclosed to commercial third parties vary across products and are rarely communicated to older adults or their families in terms that allow informed consent. The AllSeniors guide on AI companion care for 2026 is explicit about what AI companions cannot guarantee: they cannot prevent falls or emergencies, cannot replace clinical care, may misunderstand speech or miss context, and may create privacy concerns if recording, monitoring, or caregiver dashboards are not clearly explained.
The scam vulnerability dimension receives insufficient attention. Older adults with cognitive decline, social isolation, or reduced digital literacy are specifically vulnerable to AI systems — whether maliciously designed or poorly governed — that exploit emotional connection, simulate trust, and extract financial information or decisions. Agentic AI systems that manage older adults’ digital interactions, including financial communications, represent a significant abuse vector that is not adequately governed by existing elder financial abuse frameworks, which were not designed for AI-mediated exploitation.
The Substitution Risk
The most systemic risk in AI elderly care is substitution: the use of AI technology to reduce human caregiving investment rather than to supplement it. If AI companions, monitoring systems, and health AI are deployed as replacements for human contact, clinical assessment, and caregiver presence rather than as supplements to them, the net effect may be that older adults receive more technology and less human connection — which is not an improvement for a population whose primary deficit is social isolation and whose care needs require human judgement that AI cannot provide.
The AllSeniors guidance captures the right framing: AI may help remind a senior to connect, but it should not become their only connection. A better care plan includes real contact: family calls, visits, community programmes, senior events, faith communities, transportation support, and care coordination when needed. AI elderly care technology is most beneficial when it supports human caregiving by extending its reach, improving its targeting, and filling gaps in availability — not when it is deployed as a substitute for the human care that older adults need and deserve.
For readers following AI in healthcare, LiveAIWire’s analysis of AI therapy evidence and our coverage of AI in mental health addresses parallel questions about when AI health technologies help and when they fall short of their promises. Our examination of AI companion research explores the emotional and developmental consequences of algorithmic companionship in more depth.
The Cognitive Stimulation Evidence
Among the AI applications in elderly care with the strongest independent evidence base is cognitive stimulation — the use of AI to provide mentally engaging activity that may slow cognitive decline. Social robots with emotional AI capabilities can provide conversation, games, memory exercises, and responsive social interaction that most care environments cannot sustain through human staffing ratios alone. The Frontiers in Aging February 2026 review found that social robots with emotional intelligence — using affective computing to evaluate or predict a person’s emotional state — can enhance cognitive function and emotional wellbeing in elderly care settings and help combat loneliness.
The systematic review of AI applications to reduce loneliness among older adults published in Healthcare in February 2025, covering studies published through January 2024, found that while the evidence base was growing, the quality and comparability of studies varied substantially. The most rigorous interventions showed genuine short-term benefits in self-reported loneliness and cognitive engagement. The least rigorous showed larger apparent effects that the review authors attributed to methodological limitations. The honest summary is that AI cognitive stimulation for older adults shows genuine promise in clinical research, and the mechanisms by which it might help are theoretically coherent. The evidence is not yet strong enough to prescribe specific products or approaches with confidence, and the best implementations involve AI tools as supplements to rather than replacements for human engagement and professional oversight of cognitive health.
The Practical Guidance for Families
For families navigating elderly care technology decisions, the AllSeniors 2026 guidance captures the key practical framework: AI can provide reminders, conversation prompts, fall alerts, and health monitoring — but it cannot guarantee safety, replace clinical judgment, or substitute for human relationships. The right questions to ask before deploying any AI elderly care technology are: What specific need is this tool addressing? Has it been clinically validated in a population similar to our family member? What data does it collect, how is that data protected, and who has access to it? How does the person being cared for feel about using it? And what human care and contact does this tool supplement rather than replace? Technology that meets those criteria — addressing a specific need with validated evidence, transparent data practices, the cared-for person’s informed engagement, and positioning as a supplement to human care — can make a genuine contribution to extending independence and improving quality of life. Technology that fails those criteria is at best an expensive placebo and at worst a risk.
The Caregiver Workforce Crisis That AI Cannot Fully Solve
The demographic driver behind AI elderly care investment is a caregiver workforce crisis that is structural and worsening. In the United States, the Bureau of Labor Statistics projects a need for 1.2 million additional home health and personal care aides by 2030, against a labour supply that is not growing at the required rate. In Japan, South Korea, and several European countries, the ratio of working-age adults to elderly dependents is deteriorating faster than any combination of immigration policy and workforce participation rate can compensate for. AI is being deployed, in part, as a response to this structural gap — extending the reach of a shrinking caregiver workforce by handling monitoring, medication management, and social interaction functions that human caregivers previously provided in person. That deployment is rational given the alternative of unmet care needs. But it is worth being clear that AI in elderly care is partly a response to a political failure to invest adequately in human caregiving workforce development and compensation, and that the AI tools being deployed as supplements to inadequate human care are not equivalent substitutes for that care. The risk is that AI elderly care technology becomes a justification for continued under-investment in human caregiving rather than a genuine supplement to it.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, emerging technology, and their impact on business, society, and everyday life. LiveAIWire publishes original AI journalism every weekday at liveaiwire.com.