Non-Wearable Fall Detection: How the World Is Rethinking Home Health Monitoring

From panic buttons to AI that sees

If you pictured "elderly safety technology" ten years ago, you pictured a pendant: a button worn on a lanyard, pressed in an emergency to call for help. It defined the category for a generation. Around the world, it is now being replaced by something fundamentally different, and much more capable: artificial intelligence that recognises a fall on its own, by understanding how a human body moves.

This is the same class of technology that powers self-driving cars and medical scanners, applied to one job in the home: spotting the moment a person falls, or stops moving when they should not, and raising the alarm in seconds, without anyone having to press anything. The best of these systems are built privacy-first, so they protect a person without ever watching them in the way a security camera would. That is exactly the approach behind SmartGuardian.

It is also no longer a fringe idea sold only to families. As the section on government adoption below shows, national health services and city councils are now funding, trialling and rolling out this technology themselves. To understand why, it helps to understand how it actually works.

How AI machine vision actually detects a fall

"Machine vision" sounds like a camera pointed at your parent. The reality, in a well-designed system, is almost the opposite. The intelligence is not in recording a picture, it is in interpreting movement. An AI model trained on millions of examples learns the difference between someone sitting down and someone collapsing, between a normal night and a person who has gone to the floor and not got up.

Crucially, the privacy-first systems never produce a watchable video. The AI reduces every person in the room to an anonymous moving skeleton, a stick figure, and works entirely from that. There is no face, no detail, nothing a human could identify. The processing happens on the device itself, and the only thing that ever leaves the home is a simple alert.

The takeaway: the value is in the AI reading movement, not in keeping a picture. That is what lets a system catch a fall the same second it happens while protecting a person's dignity, and it is why families, and increasingly governments, are choosing it over a button on a cord.

There is a spectrum of hardware behind this idea. Some systems use a discreet infrared or optical sensor; others use radar, which detects movement through radio waves and produces no image of any kind; others use under-mattress or motion sensors. What they share is the AI layer that turns raw movement into an understanding of what is happening, and the design choice to keep a person anonymous. SmartGuardian sits firmly in this privacy-first camp: the AI sees the shape of movement, never a face, and no footage exists for anyone to view.

Why now? An ageing world, all at once

The force behind all of this is that the world is getting older, quickly, and almost everywhere at the same time. The figures from the World Health Organization are stark.

The takeaway: the number of people aged 60 and over will more than double between 2020 and 2050. By 2030, 1 in 6 people on earth will be over 60. The number aged 80 and over is set to triple, to roughly 426 million.

More older people would be challenge enough. But it arrives alongside a shrinking pool of working-age carers, and a near-universal preference: survey after survey, in country after country, finds that older people overwhelmingly want to stay in their own homes rather than move into residential care. The American Association of Retired Persons (AARP) found in 2024 that 75% of over-50s want to remain in their current home. That combination, more older people, fewer carers, a strong wish to age in place, is exactly why governments have started treating fall-detection technology as public infrastructure, not a gadget.

Why the pendant fails when it matters most

Falls are the reason any of this exists, and the scale is global. The WHO estimates that around 684,000 people die from falls every year, making falls the second leading cause of death from unintentional injury worldwide. Adults over 60 suffer the greatest number of fatal falls, and roughly 1 in 4 to 1 in 3 people over 65 fall each year.

That is the heart of the problem. A pendant only works if the person is wearing it, can reach it, and is able to press it, and a frightened, injured, confused or unconscious person frequently cannot do all three. The "long lie" that follows, the time spent undiscovered on the floor, is what turns a survivable fall into a dangerous one, sharply raising the odds of hospital admission and a move into long-term care.

Governments are backing it: trials, policy and rollouts

The clearest sign that this technology has matured is who is now paying for it. This is no longer just families buying a product. Health services, research agencies and city governments are funding pilots, changing policy, and rolling it out at national scale. A tour of recent examples shows how fast it is moving.

Spain

Valencia's camera-AI that records nothing, and Madrid's AI telecare

The clearest government machine-vision example in Spain comes from Valencia. The regional government (Generalitat Valenciana), with European Union funding, is running a pilot of an AI system called Verif-AI that watches for falls and changes in posture from a camera feed without recording any images at all, tested on the night shift in two care homes near Valencia. It is the precise model this guide describes: AI that reads movement, keeps no video, and protects residents' privacy.

Madrid, meanwhile, runs one of Europe's largest "advanced telecare" programmes. Madrid City Council has deployed around 24,000 fall detectors across older residents' homes, with an AI layer that learns each household's normal routine to flag problems early, and has begun a predictive-AI pilot in 200 homes. (Madrid's system is built on wearable and ambient sensors rather than machine vision, but it shows a major European capital treating AI fall detection as a core public service.) Nationally, Spain's public telecare is shifting to this "advanced", sensor-and-AI model, and in Andalucía alone it now reaches more than 100,000 homes.

United Kingdom

The NHS already uses camera-AI to catch falls

In England, an AI system called Oxevision is used across roughly half of NHS mental-health trusts. It uses an infrared, camera-based sensor and AI to monitor patient safety, measure pulse and breathing contactlessly, and flag falls, without giving staff a continuous, watchable video feed. An evaluation across five NHS trusts reported a 48% reduction in bedroom falls on older-adult wards, along with large drops in fall-related A&E visits. Britain is also forcing the wider shift: the national switchover of telecare from old analogue phone lines to digital (now due by early 2027) is pushing every council and care provider toward smarter, data-capable monitoring, and in 2025 NHS England began a nationwide rollout of an AI tool that predicts and helps prevent falls in home care.

Japan

Policy that rewards "watch-over" technology

No country is older than Japan, and it has turned to technology as national policy. Two government ministries jointly name "monitoring" as a priority field for care technology, the national research agency funds its development, and since a 2021 reform of the long-term care insurance system, care homes that use watch-over sensors are allowed to adjust their night-time staffing. Systems built for this market, such as Konica Minolta's HitomeQ, use a ceiling-mounted near-infrared camera and AI to detect when a resident gets out of bed or falls. By 2020, monitoring technology was already the most common type of care robot in Japanese nursing homes.

United States

Government-funded research, peer-reviewed results

In the United States, the clearest machine-vision example is SafelyYou, a camera-and-AI fall-detection system whose development was funded in part by the US National Institute on Aging. It is now used across 34 states. A peer-reviewed study of 11 memory-care communities found 41% fewer falls and 69% fewer fall-related emergency-room visits. Its privacy model keeps video only in the short window around a detected fall, for staff to verify what happened, rather than recording continuously.

Singapore

A national rollout, deliberately camera-free

Singapore shows the gov-tech version of this idea. The Housing & Development Board, which houses most Singaporeans, has piloted and rolled out in-home fall detection, with the Ministry of Health and the Agency for Integrated Care subsidising the cost by up to 80%. Tellingly, Singapore chose privacy-preserving radar and similar sensors rather than cameras, so the system detects a fall while producing no image at all, the same privacy-first principle, reached by a different route.

And beyond

The pattern repeats. In the Netherlands and Belgium, computer-vision systems (Kepler Vision's "Night Nurse" and Nobi's smart ceiling lamp) detect falls in care homes, with public research backing. South Korea launched a national AI-care strategy in 2026 built around in-home and in-facility monitoring. Australia, after its Royal Commission into aged care, funded a national project with its science agency CSIRO to predict falls from ambient sensors. Across the European Union, meanwhile, the new AI Act and long-standing data-protection law are drawing the rules of the road, and they favour exactly this design: a system that simply detects a fall, without identifying who a person is and without keeping footage, sits comfortably on the right side of those lines. Privacy-first is not just kinder. It is becoming the compliant default.

How big, and how fast, is this growing?

The investment is moving the same way as the policy. The dedicated fall-detection systems market was worth around $447 million in 2023 and is forecast to reach roughly $748 million by 2030 (Grand View Research). But that narrow figure understates the story, because the real growth is in the broader ambient and remote-monitoring categories that AI fall detection sits inside.

The takeaway: the broad ambient-monitoring category is growing several times faster than the traditional fall-alarm market. Capital and public policy are both pointing the same way, toward the home, and toward AI rather than buttons.

Remote patient monitoring is projected to approach $57 billion globally by 2030, growing at about 13% a year (MarketsandMarkets), and the wider ambient assisted living market is forecast to grow at nearly 27% a year (Grand View Research). When the demographics, the investment and the regulation all line up behind the same idea, it stops being a trend and becomes the new standard.

What this means for Ireland

Ireland sits squarely on the same curve as every country above. The population is ageing, the stated national preference is firmly for ageing at home, and the evidence that pendant alarms fail when they are needed most applies here exactly as it does in Cambridge, Valencia or Tokyo.

What has been missing is an Irish provider bringing this same technology, properly, to Irish homes. That is what SmartGuardian does. It delivers the privacy-first AI now being trialled by health services and governments worldwide: fall detection with no wearable and no button to press, working in the rooms where falls are most likely and most serious, including the bathroom, where a person would rarely be wearing a pendant anyway.

And it answers the obvious worry head-on. SmartGuardian is built privacy-first: the AI watches movement and converts it to an anonymous stick figure, processed on the device, with no video kept, transmitted or shown to anyone, including us. That is precisely the design that the best international examples, and the EU's own rules, are converging on. SmartCare Living is not importing a novelty. It is Ireland's version of a shift already well under way across Spain, the UK, Japan, the US and Singapore. You can read more about how this fits the Irish context in our guide to AI-powered home care in Ireland.

The bottom line

For thirty years, the question families asked was "will Mum press the button?" The honest, evidence-based answer was too often "no". The world has now landed on a better one: let the home itself recognise the fall, with AI that reads movement rather than watches a person, so that no one has to wear, charge, or remember anything.

From an NHS ward in England to a care home in Valencia to a public-housing flat in Singapore, the direction of travel is unmistakable, and it is being set by governments as much as by companies. Privacy-first, non-wearable AI fall detection is becoming the standard way to help an older person live safely and independently at home. Ireland is part of that story, and so is your family if you choose it.

Next steps

If you would like to talk through what privacy-first AI fall detection could look like for your own family, our team offers a complimentary 10-minute callback. We will give you an honest recommendation, including whether a SmartGuardian system is the right fit or whether simpler changes would do the job.

You can also take our free 2-minute home safety assessment for a personalised recommendation, or read our companion guides on why passive monitoring beats wearables and how to prevent elderly falls at home.