The robots are coming. And, really, if we’re honest with ourselves, they are already here. On some level, this is a semantic issue. How do we define robots? Are they humanoid machines or AI programs or automated processes? Or a combination thereof? How many of us have invited Roombas and Alexa and Siri into our lives — intentionally or without having thought about it much at all?
Regardless of how we define them, popular media, industry, and academic research have made much of their potential for caring for our aging population. Some 17% of the world’s population will be over the age of 65 by 2050, according to the UN Department of Economic and Social Affairs. If these reports are to be believed, robots may offer support that the younger population is increasingly unwilling or unable to provide. We have seen images of robots cradling frail elderly people and heard stories of robot toys keeping them engaged. Other robotic solutions fall in between, serving purely physical needs such as lifting or engaging with people on cognitive and emotional levels.
Reactions to these supposed innovations vary. Some instinctively find them vile and dehumanizing, replacing duties that they instinctively feel ought to be fulfilled by other human beings. Others are encouraged by the possibilities for refinement of the deeply flawed eldercare industry, fraught with abuse and neglect — and relieved by the potential for a stopgap when population replacement levels are falling in developed countries.
Experts in the robotics field and ethical observers are attentive to these concerns but give them varying weight. Some contend that eldercare robots are largely a fiction, a gambit that gains attention and research funding for technology that will ultimately be deployed elsewhere. Others acknowledge that these robots may fall short in some ways but improve elders’ lives in others.
Advocates are particularly enthusiastic about the possibility that robots may prolong the ability of older people to “age in place” — remaining in their homes even as their faculties decline. Those suffering from mild cognitive impairment (MCI) and physical disability may benefit from machines picking up the slack, both in terms of physical labor and the facilitation of mental engagement. Further, they argue, these machines may allow actual humans to spend more time on caring tasks by freeing them from manual labor such as cleaning.
Whether elders are interested in having robots in their lives remains an open and essential question. So, too, the wide range of abilities and needs of a diverse aging population will be key in how robotics will be implemented in their care.
“I see elder care as an enormous area for robotics going forward — allowing people to live independently for longer and age in place,” says Cynthia Matuszek, assistant professor at the Interactive Robotics and Language Lab at the University of Maryland. But she cautions, “You can’t have robots that help with that if you can’t, as a senior, get them to do what you want them to do.” She thinks language processing is one way of doing so, but robotics may assist in other areas, too.
Regardless of how one conceives this burgeoning technology, it appears to be profitable — it was valued at $2.5 billion last year by one analytics firm. Governments have contributed millions to the investments in eldercare robotics technology in anticipation of the growing needs of their aging populations.
Here, InformationWeek investigates this complex and ethically fraught landscape, plumbing the literature and including insights from Matuszek, Robert Sparrow, professor of philosophy at the Monash Data Futures Institute, and Matt Beane, assistant professor in the technology management program at the University of California, Santa Barbara.
An Aging Population
Advocates of eldercare robotic technology frequently cite the urgency of the increasing population of older adults — and the declining population of younger people who might care for them. Robots, they contend, will be essential in making up the shortfall. Indeed, in many developed countries, the aging population is rapidly outpacing birthrates.
In Japan, where eldercare robots are increasingly touted as a solution, birthrates have precipitously declined since the 1970s, with a record low this year. Meanwhile, the aging population — those over 65 — is nearing 30% of the population. That is around 1 in 10 people. As a result, the country has been characterized as the leader in a growing number of superaged societies.
Japanese culture has been particularly receptive to robotics as a solution, in part due to government programs to promote the idea and the prevalence of animist beliefs that imbue inanimate objects with a sort of life force.
Japan is hardly alone in failing to match population replacement levels. In the United States, elders now make up 18% of the population and are projected to make up nearly a quarter by 2054. It is on track to join Japan in the superaged club by 2030 — adding to the more than 30 other countries estimated to enter superaged territory that year. Simultaneously, birth rates continue to drop, with another record low reached in 2022.
Cynthia Matuszek
The growth of the aged population is a global phenomenon. Some 10% of the world population is now over age 65. It will become more acute in the coming decades. By 2050, some 16% of the world population may fall into that age bracket.
Even if paid caregivers can step into the gap and in some cases fill the roles that until relatively recently were filled by younger family members, usually women, they will not fill it entirely. One estimate suggests that in the United States alone, there will be a shortfall of some 151,000 eldercare workers. And Japan may see even greater deficits — 380,000 — as early as next year.
Practical Implementation of Eldercare Robots
Perhaps the most well-known image of robotics in eldercare is one depicting Robear, a model rolled out in 2015. It cradles a researcher in its arms, suggesting an ability to care for and protect elderly patients. The thing is it was never implemented. And despite its friendly face, Robear had limited function — it is simply an assistive model, meant to help physically disabled patients to their wheelchairs or to the toilet. Whether it could ever have provided much assistance to elders without substantial human aid seems dubious.
Despite utopian prognostications of robots taking over the laborious tasks involved in making sure that older people are properly cared for, they are largely absent from most home and institutional settings today. While breathless news stories have indicated that some care homes use upwards of 20 robotic models, a 2019 survey found that only 10% had begun using any type of robotic technology. A few years later, less than a quarter had done so — and most of the robots adopted were glorified surveillance technology, designed to detect falls, for example.
Robots are very uncommon in United States nursing homes and, in the EU, only a few models have been implemented. GoCart, a robotic transfer system that can deliver meals, has been deployed in some places. Similar robots made by Labrador Systems were adopted by a handful of American eldercare facilities in 2022. But they are far from standard technology. In fact, it is difficult to find evidence of most of these robots in action today.
Though the Danish government reportedly purchased 400 models of the robotic seal Paro, a companion animal robot, a researcher who visited one home found that they were not in use. The paper includes an image of a crusty Paro model, apparently asleep.
Still, the Japanese government has invested millions into programs that will supposedly develop more advanced and useful eldercare robots. As of 2018, it had invested upwards of $300 million in the technology. Between 2015 and 2020, the European Union dropped some $103 million on its Robotics for Aging Well program. And in 2019, the government of the United Kingdom announced that it had allocated $48 million to research on care robots. The year before, Pepper, a care robot, appeared before the UK Parliament.
But many of these robots have proven impractical, ending up shoved to the back of closets after the initial novelty wore off. Even the lower-impact care animal models seem to lose their appeal after a short period. They can prove problematic in other ways as well. In one case, an elderly man kept trying to remove the artificial fur to discern how the robot worked. And a woman with dementia became overly attached, refusing to go to bed without the pet robot.
Further, staffers quickly grow weary of the additional work involved in ensuring that they are operating properly.
“Non-managerial staff are the people who silently take up that burden,” Beane says.
Matt Beane; Photo credit Jon Chomitz
Maintenance is a major issue, given the proprietary nature of these creations. The companies that manufacture them must be contacted in cases of malfunction — spare parts are largely unavailable on the open market and there are no repair technicians on call to get things up and running in a matter of days. “You have to wait for a service rep to show up at your location, get in there, take it apart, put it back together,” Beane says. If you think ‘right to repair’ is an issue with cell phones and computers, think about the implications for sophisticated, purpose-built robots.
And then there is the matter of upgrades. “The update to a robot is not just digital. It’s not like you have to learn this new set of prompts or something. It has new hardware. You have to plug it in differently. You have to clean it differently. It needs new parts. It’s just a whole multidimensional paradigm shift,” Beane explains.
This begs the question as to whether these robots were ever intended to be useful anyway, Sparrow thinks. “We need another compelling social narrative — aged care robots. They trot out this sort of looming demographic crisis — millions of lonely old people. And then they start building robots that they were going to build anyway,” he says of the industry.
“Machines simply can’t walk into your apartment and do something useful,” he adds. “We can’t build those machines at the moment.”
Matuszek agrees but is encouraged by more specific applications. “There’s some incredible work in feeding robots that help people who have limb difficulties — just getting food into their mouth,” she says. “There’s some incredible work in prosthetics. There’s work in social robots — getting people to engage, not be too isolated.”
“We’re going to see robots that are good at progressively more things — a robot that can load the dishwasher, clean up the counters and stir the pots,” she predicts. “We’ll see sort of these capabilities getting added in over time. But as for challenges — boy, there’s a lot. It’s hard stuff.”
There will be further challenges in directing robots to complete tasks effectively. Natural language processing will need to advance significantly — the ability to understand partially formed instructions or indirect language will be essential.
Matuszek thinks that such innovations may be helpful in facilitating aging in place — allowing people to avoid institutionalization as long as possible. Adjustments to a home that might seem expensive on their face value may be more cost-effective than putting an elderly person in a care home. She suggests that such renovations as lowering shelves and installing elevators may, in combination with care robots, ultimately be cheaper.
Eldercare Robots for Physical Tasks
While implementation remains rare, researchers remain committed to robotic eldercare technology. Among the most promising developments are in the realm of physical assistance. Physical disability is often one of the major factors that lead to elders being transferred to assisted living facilities — which may lead to declines in their mental health as well. Physical assistance may allow them to live on their own for longer, facilitating a sense of autonomy and bolstering mental well-being.
And in assisted care contexts, the completion of physical tasks by robots may ultimately offer relief to care workers as well. The physical strain of assisting disabled elderly people has been well-documented — rates of injury are high, with many enduring back strain from repeatedly lifting their charges.
Robots such as Robear and Hug were intended to assist in lifting patients for transfer to other locations. “They’re extraordinarily expensive, hard to maintain, hard to reposition, and create a lot of extra work for the people around them. It’s a profoundly interesting demo, but it’s not a practically serviceable technology,” Beane says. While neither appears to have been adopted, researchers are working on enhancing this type of technology.
One group aims to create a sit-to-stand robot that will not require the presence of a human assistant, with grab bars that adjust to the user’s needs and allow the machine to serve as a sort of mechanical walker. It has been suggested that repurposing devices designed to ease physically demanding labor, such as the wearable HAL Lumbar Type for Labor Support, may prolong physical independence. Even more simplistic technology such as Resyone, a bed that transforms to a wheelchair, may be helpful. “Roboticists are prone to adding the things that are possible because they like to solve cool technical problems,” Beane observes. But in the near term, it is likely that these more parsimonious solutions will be the way that robotic technology makes its way into eldercare.
Others take more Alexa-like forms — reminding older people to take medications, for example, or to have a meal. Others serve to detect falls, which can result in major health complications.
Delivery systems assist in retrieving items, thus decreasing strain and reducing the risk of injuries. Some are essentially motorized carts, as in the case of GoCart and Labrador Systems technology, called the Retriever Pro. “That’s a very elegant solution to a mobility problem and an object acquisition problem around the house. It is designed in such a way that it is unobtrusive, behaves in a very focused and subservient way,” Beane says.
Others, such as ADAM, serve similar functions but are more humanoid and elaborate. ADAM boasts mobile, dexterous arms and a visual perception system, allowing it to navigate obstacles. It can complete simple household tasks such as sweeping the floor or setting the table.
In one setup ADAM was able to communicate with the RobWell, another robotic system that uses bodily monitoring to predict such requirements as food and water and then deliver them as needed.
A wide variety of other robotic systems have been designed to offer healthcare monitoring and physical assistance as well — from KOMPAÏ Assist, which offers medication reminders and mobility help to RAMCIP, which can execute a wide range of tasks, from detecting gas leaks to retrieving items that have fallen.
Eldercare Robots for Emotional and Cognitive Tasks
While robots that take up the slack of physical labor for elders and their caregivers have shown the most immediate promise, a variety of technologies that offer solutions to the emotional and cognitive challenges of aging have been proposed and trialed as well.
Paro, the robotic seal, may be the most well-known example. It looks like a stuffed animal but is equipped with a robotic system beneath its artificial fur. It is meant to emulate a companion animal, making sounds and reacting to touch. Aside from reducing stress and offering a point of focus, companion animals may even offer cognitive benefits by keeping elders engaged, with studies even showing better memory retention among older pet owners. Robotic animals may offer some of the same benefits. There is some indication that these types of robotic animals even ease sundowner’s syndrome, in which dementia patients become more agitated during the evening hours.
“We don’t have enough money or time or human beings to have somebody provide near-constant interaction and attention and care to somebody the way that a pet can,” Beane says. “Go get the data on somebody with dementia who’s alone for six hours — how their life goes, what happens to their health, what their subjective life experience is. It’s easy to make assumptions about how much better that person would be without that seal in the room.”
Other, more humanoid robots, claim to offer further benefits — from reminding elders to eat and take medication to walking them through exercises or playing games.
One experimental robot, Casper, attempts to guide elders experiencing cognitive decline by speaking to them and guiding them through tasks, sometimes using videos displayed on an iPad embedded in its chest. The robot features a glowing, human-like face, which its designers intended as a way of facilitating emotional attachment. It can converse with its human counterparts, play games with them, and remind them of essential tasks. It is mobile and equipped with an array of sensors and scanners that help it navigate the environment and react to human behavior.
Similarly, Pepper, a general use model proposed for use in professional and domestic tasks, can offer reminders and guidance. Another model, Stevie, aims to do the same. These types of robots can also guide seniors through physical activities and facilitate conversations with family members and medical care providers.
Others, such as the discontinued Zora Bot program, attempt to disrupt behavioral disturbances such as wandering — a major problem for people who experience cognitive decline — by detecting those disturbances and then engaging and redirecting the person.
Robert Sparrow; Credit Tatiana Koulatsos, Red Butterfly Photography
While a range of encouraging studies have indicated that these robots — at least experimentally — have made improvements in the lives of seniors, Sparrow questions their methodologies. They don’t typically control for the effect of the observers or for the novelty of the new devices, he says. “When you bring a new robot into a setting, everyone says, ‘Have you seen the robot? And what did the robot do?’ But if you bring in a plant or a new painting, people also have new conversations which impact their wellbeing. They bring in a robot, they see some change in behavior, and they say, ‘Oh, it’s because it’s a robot.’ But it might just be because it has disrupted the routine,” he suggests.
Regardless, most of these models likely have a way to go before they can be implemented at scale.
And elders themselves remain skeptical. One study that surveyed perceptions of Pepper in dementia care found that the robot was off-putting and that many respondents did not believe it could be trusted.
Ethical Implications of Eldercare Robots
The ethical concerns underlying the entry of the robotics industry into eldercare are substantial. For one, the reasoning behind using robots to care for elders — many of whom are loved and respected — can be interpreted as highly manipulative. Caring for seniors is challenging and often very draining, both emotionally and financially. The idea that mechanized labor may be able to step in and alleviate the strain is incredibly appealing. That appeal may be leveraged by the industry to encourage adoption of untested technologies, the consequences of which are unknown.
“The organization wants it to look like a good bet,” Beane says of institutions that adopt these robots. “They will publish news stories about their robot. If it’s not working out, the organization will sort of tuck it away but keep it on the website.”
There is significant motivation to push these technologies in societies like Japan, where they have been proposed as an alternative to immigrant labor. Thus, mechanized care technology has ideological and even governmentally subsidized motivations that may not take the best interests of the actual recipients of care into account. Often, these technologies do not actually include elders in their development processes and only study their effects in tepid, poorly designed trials after the robot itself has been constructed. This is a known problem in engineering in general — the failure to consider the effects on the end user.
Safety must also be considered — robots may malfunction, misdirect users, or even present mundane challenges such as fall hazards. What happens if an elderly person trips over a care robot or one of them topples onto them?
Surveys of older people suggest substantial privacy concerns as well. A focus group respondent conveyed her distrust of eldercare robots, saying “It must not be a blabber. After all, it is there to keep the secret.” While monitoring of medications, lifestyle habits, and even potential falls may be helpful to relatives and caregivers, elders remain concerned about autonomy and control of their private information.
“Older people are not demanding robots,” Sparrow notes. “Engineers want to build robots. And to my mind, that is itself problematic. We are offering people a ‘solution’ that they don’t want. What people want is to be included in the community, to remain in their own homes. They don’t want to be entertained by some toy. They don’t want to lose access to human support services.”
Indeed, the assumption that the emotional requirements of elderly people may be met by robotic interaction — rather than simply augmented — remains a concern. Will the emotional burden of supporting people with declining physical and mental faculties ultimately be offloaded to robots simply because it is easy to do so? And will these machines adequately fulfill the highly complex suite of human emotional needs?
“It’s just down to good design — trying for something other than simple one for one replacement, trying to fill the cracks where there’s a gaping hole,” Beane argues.
Ethicists further raise the problem of deception — if the person tended to by the robot actually believes that the machine cares for them in the way that a human does, the scenario may constitute an elaborate lie.
“The alternative isn’t usually a dedicated human who loves you and is happy to be there,” Matuszek counters. “The alternative is frequently isolation.”
So, it seems the eldercare crisis may come down to a set of suboptimal choices: relying on an increasingly strained human system, implementing suboptimal robotic technologies, or living in the hopes that more ideal solutions will emerge.
