Many a roboticist claims to be building their robot to “take care of the elderly”, but is that genuine altruism, or just a flimsy excuse? In my country of the Netherlands, half of all robotics research seems to be focused on healthcare applications. As in Japan, Dutch society is trying to handle the trend of “greying”: The growth and longevity of the elderly population, combined with lower birth rates, results in more elderly than young people available to take care of them. My parents, pensioners themselves, help out at a local care home, where retirees speak excitedly about the promise of healthcare robots, but also doubt whether technology will pull through for them.
In this series of articles I want to explore how robots can realistically be of help to the elderly population. As elderly care is a very broad topic, I will divide it into three categories: Robot pets, assistants, and nurses, after which I shall further elaborate on the realism of adopting these technologies. In this first article, we will examine the pros and cons of the most prominent robot pets.
Paro the therapeutic robot seal was designed to provide emotional healthcare rather than physical. The robot is targeted at people with dementia and Alzheimer in nursing homes, who are difficult to handle when becoming confused or agitated. The baby seal’s cute interactive behaviour invites stroking its fur, which has a known calming effect that lowers blood pressure, and its presence facilitates connecting with other people socially. Real pets can fulfill the same role, but require a dependable caretaker, and are often not allowed in nursing homes (Another debate worth having).
The form of a white seal was chosen after trials revealed that people had considerable preconceptions about cat and dog robots that could not be satisfied. Although the robot is still clearly not a real animal, patients with dementia often don’t realise or mind whether it is. It provides emotional stability to most patients, though not in every case.
When I asked a local professor involved in Paro’s deployment about the concern whether a personal robot pet could make family feel less needed to visit, she explained that the Paro robot was actually so expensive ($6000!) that nursing homes could only afford to buy one and time-share it between patients, which rules out that scenario. The price tag severely limits how many elderly Paro will be helping.
At the other end of the financial spectrum is a therapeutic robot cat by toy maker Hasbro, that only costs $125. Petting the cat triggers movement, meowing, and gently vibrating purrs. It has the same benefit as Paro: Stroking fur relaxes people, and the interaction instills a sense of companionship. Research on this toy corroborate these results, and customer reviews are overwhelmingly positive, with the majority gifting the robot to 90-year-old parents with dementia who used to have a cat, but aren’t allowed real cats in nursing homes.
The illusion of life is less convincing than Paro’s: The robot cat does not look as realistic as we know cats to look, is limited to a lying position, meows cartoonishly, and feels hard to the touch, but many dementia patients still take it for real and sometimes become concerned why it does not eat or drink. A few customers have reported the robot’s mechanical lifespan as only one year, but given the choice, I would try a $125 cat before a $6000 seal. There also exists a dog version, but I find neither its constant sitting pose nor its frequent barking as convincing as a cat that would definitely lie down all day.
Sony’s Aibo dog is marketed as a companion robot, and has a great deal more interaction and animation than its stationary competitors. The few studies related to dementia patients describe many of the same positive effects, but are unconvincing. Some studies lasted only 3 days, which does not take the novelty factor into account. Most research on Aibo was also conducted in Japan, which has a particularly welcoming culture towards robots. The only other country where Aibo is currently available is the USA, where it costs a pricey $2900.
With Aibo’s clearly robotic design, it does not provide the stress-reducing effect of fur, and people have no natural familiarity with it. Because of this, I do not find Aibo suitable as an intervention device for agitated dementia patients. It may however have a modest role as companion. Many adult Aibo owners find Aibo’s advanced behaviour convincingly alive, and consider it part of their family. Aibo’s latest (2019) autonomous patrol feature certainly adds to that experience, but also makes it a tripping hazard for fragile elderly. To mitigate this, the robot plays a toyish melody whilst roaming the room.
Judging from relatively shallow tech reviews, I expect that Aibo’s reception will strongly depend on a person’s willingness to anthropomorphise, and isn’t everyone’s cup of tea. One important consideration is that after the first three years, failing to extend the subscription to Sony’s online services with a yearly $300 will result in complete memory loss, after which Aibo will no longer have its developed personality, recognise faces, or respond to commands. There is also no telling when Sony will unilaterally pull the plug on Aibo as they did in 2006 due to disappointing sales. A robot that itself can suffer dementia does not seem ideal.
Having read the results of several long trials and studies, my initial concerns have been settled: Pets, robotic or otherwise, slightly reduce loneliness, improve overall mood, and actually facilitate family visits rather than replace human contact. Especially in cases of Alzheimer, it is much more comfortable for all to socially connect over a pet, than it is for the patient to regard their children as unwanted strangers. For nurses and patients both, it is also much more preferable to pacify a rowdy patient with an interactive toy than with a wrestling match. Most Dutch nursing homes now own one or two robot pets. Robotic pets do not reduce the workload of nurses, but they improve the quality of life.
Although robot pets are mainly employed for the niche of elderly dementia patients, I can also imagine them lifting the spirits of some elderly who are well aware that they are robots, when real animals are not an option. My grandfather might not have died so miserably in a nursing home after being separated from generations of feline familiars, if he’d had at least one familiar snout around to keep their memory alive. Of course, real animals remain preferable whenever possible, as they are far better judges of mood, offer companionship, liveliness, and routine. Robot pet developers can still learn a lot from real pets.
There remains a question on the ethics of enabling dementia patients to believe that a robot pet is a real animal when they can not tell the difference, but I would opine that such patients already live in their own world, and the feelings of joy are real, even if they are stimulated by a stuffed toy. Considering that we’ve all been children, I don’t think this is such a controversial statement. The most agreeable advice I’ve come across is to let the patient judge for themselves what it is and whether they like it.
In part 2 of this series, we will examine robot assistants.