Key takeaways
- In 2013, when about one in three people in Japan said they knew what care robots were, 65.1% said they would want care robots used for their own care.
- The appeal most chosen in that survey was a lighter physical and mental load for the carer (63.9%).
- Japan's care plans need about 570,000 extra care workers by fiscal 2040, on top of the 2.15 million working in fiscal 2022.
- In autumn 2025, 47.2% of the residential and live-in settings that answered a government survey had monitoring sensors and 56.4% had care record software or similar tools; lifting devices were in 15.8%.
- Three questions help you judge any care technology: whose work it makes lighter, who it alerts and how consent was asked, and whether it is easy to use and affordable to keep.
- —In 2013, when about one in three people in Japan said they knew what care robots were, 65.1% said they would want care robots used for their own care.
- —The appeal most chosen in that survey was a lighter physical and mental load for the carer (63.9%).
- —Japan's care plans need about 570,000 extra care workers by fiscal 2040, on top of the 2.15 million working in fiscal 2022.
- —In autumn 2025, 47.2% of the residential and live-in settings that answered a government survey had monitoring sensors and 56.4% had care record software or similar tools; lifting devices were in 15.8%.
- —Three questions help you judge any care technology: whose work it makes lighter, who it alerts and how consent was asked, and whether it is easy to use and affordable to keep.
Picture a bedroom at night. A large white robot with a friendly bear face slides its arms under an older woman, who is awake with her hand resting on its arm, and lifts her from her bed into a wheelchair. On the mattress she has just left, a small light blinks once.
The robot in that picture is real, and it was built in a laboratory. ROBEAR is an experimental care robot that RIKEN, a Japanese research institute, and the company Sumitomo Riko showed in 2015, built for one of the most strenuous jobs in care: lifting someone from a bed into a wheelchair, which RIKEN says care staff do an average of 40 times every day. [6] So, would you let a robot such as ROBEAR care for you, or for someone you love?
In August 2013, Japan's Cabinet Office asked 1,842 adults, face to face, something close to that. If you yourself came to need care, would you want care robots used? 65.1% said yes, about two in three, and 29.3% said no. At the time, 31.9% said they knew what care robots were; the rest had only heard of them or had never heard of them. [3] It is the latest national survey on the question we found, so we give its year with every figure from it.
The question is pressing in Japan. On 1 October 2025, 36.22 million people there were aged 65 or over, 29.4% of the population, and 21.27 million were 75 or over. By 2070, 38.7% of the population is projected to be 65 or over, one person in 2.6. [1] Adding up the prefectures' plans, Japan needs about 2.72 million care workers in fiscal 2040, against about 2.15 million in fiscal 2022: about 570,000 more. For fiscal 2026, the estimate is about 2.40 million. [2]
To find out what Japan's care homes really use, we read primary sources: the Cabinet Office's 2013 survey and its 2026 report on the ageing society, trials and surveys from the Ministry of Health, Labour and Welfare (MHLW), a maker's product page, the research institutes' own pages, one controlled trial, and the researcher James Wright's own summary of his book on care robots in Japan. The Japanese quotes are our translations. The newest survey is a draft summary that MHLW presented to a government committee in February 2026: 6,876 care providers replied, 35.3% of those asked, and we say "homes that answered" whenever we use it. Its figures cover residential and live-in care, meaning care homes and services where people stay overnight. [12]
This article is the written companion to the Automation Switch film Japan's Care Robots: Would You Let One Look After You?, which tells the same story on screen.
What did people in Japan hope a care robot would do?
The 2013 survey asked all 1,842 people what attracted them to care robots, and let them pick several answers. The top answer, at 63.9%, was that the physical and mental burden on the person doing the caring would get lighter. Next, at 41.5%, came a quieter wish: you would not have to feel you were imposing on the person caring for you. Third, at 35.8%: the person receiving care could do more for themselves. [3]
The survey also asked the 696 people who had cared for a family member at home what had been hardest. Helping someone go to the toilet came first (62.5%), then bathing (58.3%), meals (49.1%) and moving someone between a bed, wheelchair, toilet or bath (48.3%). [3]
Set the two answers side by side. The jobs family carers found hardest are some of the most private moments of a person's day, and in 2013, 41.5% of people chose not having to feel they were imposing on their carer as an appeal of care robots. [3]
They were practical too. Asked in 2013 about caring for a family member at home, 59.8% said they would want to use a care robot. When choosing one, people wanted it easy to use (74.4%) and cheap (68.6%). [3] Those two wishes come back at the end of this article.
Which jobs did Japan ask robots to do?
Japan's government answers its care gap in five ways: better pay and conditions, recruiting and training a wider range of people, keeping staff and raising productivity, making care work more attractive, and welcoming workers from abroad. [2] Over the past two decades, Wright writes, one group of technologies in particular has been heavily promoted by the Japanese government: robots. [5]
MHLW and the Ministry of Economy, Trade and Industry turned that into a list. Their "priority areas for robot technology in care" (six areas and 13 items in the 2017 version) cover transfer, meaning lifting someone; getting about; the toilet; watching over people and keeping them company; bathing; and care work support, such as collecting care records. Newer areas add functional training and dementia support. The two ministries fund development and adoption in these areas. [4][12]
For lifting, ROBEAR weighs 140 kilograms, down from the 230 kilograms of its predecessor, RIBA-II, and RIKEN describes it as experimental. [6] HAL Lumbar Type, from the company CYBERDYNE, is worn by the carer: its maker describes a wearable cyborg that assists movement using the wearer's bioelectrical signals, to reduce the load on the lower back when moving someone from a bed, and waterproof so it can be used to help with bathing. [7]
Keeping people company drew another kind of robot. In 2016, the Japan Agency for Medical Research and Development (AMED) launched a large trial placing about 1,000 communication robots in care settings, choosing 19 kinds of robot as candidates. [8] We report the trial's launch only, because its results pages are no longer online.
Two other jobs on the list were watching over people and keeping care records.
What did a robot seal add in a controlled trial?
One robot made for keeping people company is PARO, a robot seal developed in Japan by the National Institute of Advanced Industrial Science and Technology (AIST). It is modelled on a baby harp seal, carries touch, voice and whisker sensors, gives what AIST calls a "cheerful response to caressing", and learns a name you give it. Leasing began in September 2004. [9]
A controlled trial of PARO ran abroad, in 28 long-term care facilities in South-East Queensland, Australia, with 415 people aged 60 or over living with dementia. Nine facilities had PARO, in 15-minute sessions three times a week for ten weeks; ten had the same seal with its robotics switched off, which made it a soft toy; and nine carried on with usual care. [10]
Although more effective than usual care in improving mood states and agitation, PARO was only more effective than a plush toy in encouraging engagement.
Those are the trial's own words, and mood, agitation and engagement over ten weeks are the only outcomes we report from it. [10]
Keeping people company is one job on the list. The appeal people chose most often in 2013 was a lighter load for carers, and lifting robots were built for exactly that.
What happened when the robots reached ordinary care homes?
MHLW ran its own trials of technology in care homes from fiscal 2020 to 2022, and published the results. MHLW's fiscal 2022 trial of wearable lifting robots reported that the share of staff reporting moderate to severe back pain fell slightly. A smaller carer who used to need a second person to help a larger resident could do it alone wearing the device. [11]
There was a time cost. Once putting the robot on and taking it off was counted, the time spent moving and repositioning residents rose slightly. [11]
James Wright spent 18 months of fieldwork in Japan studying the gap between the promise of care robots and their use. Robots that care homes bought, he writes, often seem to end up being used for a couple of weeks before being locked away in a cupboard. By his account, around 10% of care facilities had introduced such a device. His reasons: robots are often too expensive for care homes to buy or lease, and fail to deliver cost savings by saving labour. [5]
Wright's picture comes from the fieldwork behind his book, published in 2023. The government's survey of autumn 2025 shows what happened next, and it describes a different kind of machine.
What did Japan's care homes choose instead?
Go back to the light that blinked on the mattress. It is a bed sensor. Monitoring devices in care homes come as floor mats, bed sensors and cameras, [13] and they tell a carer when a resident wakes or gets out of bed. [11] Watching over people is one of the government's priority areas for robot technology in care, so a sensor like this falls within Japan's own definition of the field. [4]
MHLW's survey of care providers, run from September to November 2025, found monitoring devices in 47.2% of the replying residential and live-in care settings, up from 30.0% in fiscal 2022. Care work support devices, such as care record software, were in 56.4%, up from 10.2%. [12]
| Criteria | Share of settings with the device |
|---|---|
| Care work support (care records and similar) | 56.4% (10.2% in fiscal 2022) |
| Monitoring (sensors that alert staff) | 47.2% (30.0% in fiscal 2022) |
| Bathing | 26.2% |
| Staff intercoms and similar | 22.3% |
| Transfer (lifting) | 15.8% |
| Getting about | 4.4% |
| Communication | 2.9% |
| Toileting | 2.2% |
Among the special nursing homes that answered (Japan's main type of long-term care home for older people), about nine in ten (90.5%) had at least one device from the priority areas, a wearable device or a staff intercom. [12]
In MHLW's trials, homes with monitoring devices replaced fixed night rounds or made them less frequent, and used the alert to know when a resident woke or got up, so toileting care could be given without disturbing their sleep. [11] In the 2025 survey, 40.4% of the homes with sensors that answered had switched from fixed night rounds to visiting each room according to the resident's situation, and 35.8% made rounds less frequent. [12] So in the trial homes, help with the toilet, the hardest part for family carers in 2013, could be timed to when a resident woke.
Combining the trials across 111 facilities, the more residents had monitoring devices, the less time staff spent on direct care and rounds combined. [11] Japan's fiscal 2021 care fee revision eased night staffing standards for care homes that use monitoring devices. [11]
One residential home, among those registered for a scheme that lets homes using technology staff more flexibly, used to run a 16-hour night shift, from 4 pm to 10 am. It added a new 8-hour night shift beside it, and with the two patterns, staff can now take a nap. [12]
Lifting devices reached 15.8% of the same settings. For lifting, bathing and monitoring devices alike, the most common reason homes gave for adopting them was to reduce staff's physical and mental burden.
Put the pieces together. Wright found robots that cost too much and saved too little labour to pay for themselves; the machines that spread are quiet ones, such as care record software and the monitoring sensors that homes most often adopted to reduce their staff's physical and mental burden, the same appeal people chose most often in 2013.
Carers stay at the centre of it. Japan still plans for about 570,000 more care workers by 2040 than it had in 2022. [2]
How can you tell whether care technology is used well?
As more care homes adopt these machines, you may meet one, choosing care for someone you love or, one day, for yourself. Here are three questions to ask about any care technology, of a care home or a supplier. They are questions about the technology only; for care and health decisions, speak to the professionals involved.
- 01Whose work does it make lighter?
Ask what changed for staff after it arrived: rounds, visits, paperwork. In one Japanese care facility, staff found that for some residents the sensor alert sounds often, so they visit the room many times, and 32.9% of the homes with monitoring devices that answered MHLW's survey said they had enough to watch every resident individually [12]. Ask how many devices there are, and who they cover.
- 02Who does it alert, and how was consent asked?
Sensors come as floor mats, bed sensors and cameras. Cameras can show and record a resident's room, and about three in ten facilities with camera devices found respecting privacy difficult [13]. Ask who sees the alert or the picture, whether video is stored, and how the resident and family were asked for consent.
- 03Is it easy to use, and can the home keep it running?
In 2013, people wanted a care robot easy to use (74.4%) and cheap (68.6%) [3]. Wright saw robots priced beyond many care homes, and bought robots locked away after a couple of weeks [5]. Ask whether every member of staff knows how to use it, and what it costs to keep.
In a study of camera monitoring carried out for MHLW, 40% of facilities with camera devices said some staff did not know how to use them. For ordinary video cameras, concern for residents' privacy was the main reason facilities chose against them; for silhouette cameras, which show an outline, it was cost. Some makers recommend that facilities obtain consent, and some facilities obtain it on their own initiative. [13]
Go back to the woman in that bed. In nearly half of the settings that answered Japan's survey, the machine she would meet at night is a monitoring device, such as the small light on her mattress, which tells a carer she is awake so someone can come to her. Ask the three questions, and you can judge for yourself how well a device like that is used.
If you care for someone, or you are thinking about care for yourself, help is there. India: the Alzheimer's and Related Disorders Society of India (ARDSI), with services and city contacts at ardsi.org/services.php. Philippines: the Alzheimer's Disease Association of the Philippines (ADAP), alzphilippines.org, which runs support group meetings. South Africa: the Association for Dementia and Alzheimer's of South Africa (ADASA), +27 (0)11 792 2511, info@adasa.org.za. UK: the Carers UK Helpline, 0808 808 7777, Monday to Friday 9 am to 6 pm; Age UK, 0800 055 6112 (England), 0800 022 3444 (Wales), 0800 124 4222 (Scotland), 0808 808 7575 (Northern Ireland). United States: the Eldercare Locator, a public service of the Administration for Community Living, call or text 1-800-677-1116, eldercare.acl.gov. Everywhere else: a carers' organisation or a dementia association in your own country.
Get the next story in your inbox
One email a week from Automation Switch: the research behind each film, and what it means for the technology you live and work with.
Free, instant download
Take a free guide with you
17 Free AI Courses
Seventeen genuinely free AI courses from the companies and universities building the field, sorted into a path you can follow.
The SKILL.md Starter Kit
Where to find skills worth installing, a five-minute check before you install one, and a template to write your own.
Article SourcesShow referencesHide references
We reviewed the sources below to support the claims, pricing, and benchmarks referenced in this article.
36.22 million people aged 65 or over (29.4%) on 1 October 2025; 21.27 million aged 75 or over (17.3%); 38.7% projected for 2070, one person in 2.6.
About 2.15 million care workers in fiscal 2022; about 2.40 million needed in fiscal 2026 and about 2.72 million in fiscal 2040 (about 570,000 more); the government's five measures.
1,842 adults face to face; 65.1% yes and 29.3% no for their own care; 59.8% for family care at home; 31.9% knew what care robots were like; appeal (63.9%, 41.5%, 35.8%); hardest parts for 696 family carers; easy to use (74.4%) and cheap (68.6%).
The priority areas for robot technology in care, six areas and 13 items, set with the Ministry of Economy, Trade and Industry, and support for their development and adoption.
Two decades of government promotion of robots; 18 months of fieldwork; robots used for a couple of weeks then locked in a cupboard; around 10% of facilities; cost and labour savings.
ROBEAR as an experimental care robot from RIKEN and Sumitomo Riko; 140 kilograms against RIBA-II's 230; lifting from bed to wheelchair done an average of 40 times a day.
A wearable cyborg (the maker's word) that assists movement from the wearer's bioelectrical signals, reduces lower-back load when moving someone from a bed, and is waterproof for bathing help.
About 1,000 communication robots placed in care settings; 19 kinds of robot chosen.
PARO modelled on a baby harp seal; touch, voice and whisker sensors; "cheerful response to caressing"; learns a given name; leased from September 2004.
28 facilities in South-East Queensland; 415 participants; PARO, plush toy and usual care groups; 15-minute sessions three times a week for 10 weeks; the trial's conclusions on mood, agitation and engagement.
Wearable lifting trial (time rose slightly, back pain fell slightly, one carer instead of two); night rounds and toileting with monitoring devices; 111 facilities; the "soon" toileting alert; the fiscal 2021 easing of night staffing standards.
6,876 providers (35.3%), September to November 2025; adoption by device in residential and live-in settings (56.4%, 47.2%, 26.2%, 22.3%, 15.8%, 4.4%, 2.9%, 2.2%); 90.5% of special nursing homes; night rounds (40.4%, 35.8%); 32.9% enough devices; reasons for adoption; the two interviews.
Floor mat, bed sensor and camera devices; privacy; about three in ten found privacy difficult; 40% with staff unsure how to use them; reasons against ordinary and silhouette cameras; consent.
