Robots and care of the ageing self: An emerging economy of loneliness

Abstract

What happens when caring for the ageing population is so devalued that robots are deployed to care for our elders? We examine the growing employment of companion robots in elder care as one response to a critical labour shortage and loneliness epidemic shared across the Global North. Reflecting on interviews conducted with robot engineers, researchers, NGO care providers and local government, we examine five robots under development or in use in the UK and the USA. We ask if machines providing emotional and social care signal a diminishment of what it means to be human or if robots and automation present a promising solution to our elder care crisis. We do not evaluate the efficacy of robotic technology but identify and question assumptions concerning what it means to be human in modernity and examine companion or social robots at a moment of crisis and the substantive reorganisation of social reproduction wrought by neoliberal austerity. We end by calling for a reimagining of elder care, in which the care of our elders is radically revalued and where robots assist and support workers in their difficult and skilled labour of care.

Keywords

Companion robots elder care loneliness social reproduction ageing

In the 2021 German film, I’m Your Man, an archaeologist working at the Pergamon Museum in Berlin is compelled by a colleague to assess a humanoid robot in order to get funding for her own unrelated research. For 3 weeks, she is required to live with the robot, which has been designed to be her life partner. At first hostile to the robot, she gradually enters into a romantic relationship with him. But before the 3 weeks are through, she abruptly ends the relationship, and in her final report, she argues firmly against the suitability of robots as life partners. She judges them to be too accommodating and is concerned that humans who choose robots as life companions will lose the ability to be with other humans, who are sometimes annoying and sources of conflict as well as friendship and joy. Human relationships (ideally) are mutual and complex and demand reciprocity and compromise: a relationship with a robot orbits around a single self. A minor secondary plot in the film involves her father, who she visits with the robot on two occasions. Her father lives alone in the family home and shows signs of memory loss that suggest that living in this comfortable place soon will be unsustainable. It is never stated explicitly, but the film intimates that the handsome congenial robot would be appropriate for the father's care.

There is something both disquieting and promising about the prospect of enlisting robots for elder care as one answer to a labour shortage and an already existing crisis in seniors’ care. Does giving oneself over (or being given over) to a machine for emotional and social care signal a diminishment of what it is to be human and what it is to be cared for as a human (Sharkey and Sharkey, 2012; Sparrow and Sparrow, 2006; Turkle, 2017)? Does the suitability of robots for elder care evidence the already existing dehumanisation of those living with dementia (Herbrechter, 2021; Kamphof, 2015)? Or does the prospect of enlisting robots for elder care present a promising solution to serious (and expensive) societal problems in many countries in the Global North of providing adequate care to older people who require it?¹ How are robots being made to respond to, alter and/or maintain existing geographies of need and care?²

We draw on interviews with robotic designers in the USA, Israel and the UK, as well as UK-based researchers, local government and NGO providers of social care to describe a range of care robots in use and/or under development.³ Our focus is not so much on assessing the effectiveness of robots in delivering care (although those we interviewed do have things to say about this). We have little to offer to the important question of how the introduction of robots changes the work conditions of the care workers who labour beside them (see Carros et al., 2002; Schwiter and Steiner, 2020; Wright, 2019). Our contribution lies more in identifying some root assumptions that undergird these other assessments (of the efficacy of robotic care and potential disruptions to human work conditions). These include a cluster of assumptions about what it is to be human in modernity: the understanding that loneliness is a (costly) modern condition, that independence and empowerment are inherently desirable and that the private home is the ideal living environment in all phases of life. Robots are being put to work on a particular conception of the ageing subject within advanced capitalist societies, that is, the individualised self.

Capitalism, generations of feminist political economists have told us, works by parsing reproductive work (including care) from the capitalist economy, such that this activity is considered to be outside the economy (Federici, 1975, 2012a, 2012b; Folbre, 2001, 2012; Fraser, 2016; 2022; Meehan and Strauss, 2015; Mezzadri et al., 2022; Mitchell et al., 2004). Historically, care has been delivered by the family or community at little cost to capitalist enterprises. When this activity is commodified as waged work, as it now often is, it has been undervalued through entrenched relations of gender, race and imperial domination. Robots are being developed at a moment of crisis within this organisation of social reproduction, within contexts of neoliberal austerity and the withdrawal of public funding for social care. The funding through the Older Americans Act is, Engelhart (2021) calculates, 16% lower in real terms in 2019 as compared to 2001, and there will be an estimated shortfall of 150,000 paid care workers by 2034. Since 2020, more than 300 nursing homes have closed in the USA and an estimated 400,000 workers have left their work at long-term care facilities (Leland, 2022). In the UK, Skills for Care (2022) notes that, in 2021/22, there were 165,000 vacancies in adult social care – an increase of 52% from the following year, demonstrating an escalating and structural crisis in recruitment and retention. Social care has the highest labour turnover than any sector in the UK economy. As Engelhart notes, the argument in favour of social robots is that they ‘are better than nothing – and nothing is on its way'. Solutions once found in social welfare or social care provisions are now sought through other means. Care robots are thus an emerging face of the privatisation of commodified social care. The development and commercialisation of these robots sharpen questions of whether care of ageing persons should be a site for capitalist enterprise and profit-making and whether and how public monies undergird this ‘loneliness economy' (Engelhart, 2021).

In framing the issue thus, we are not calling on or posing a generalised ethical critique of the automation of care. We are questioning the ideological foundations of the debate. Drawing on Fraser's view that a crisis of care creates opportunities to contest capitalism more broadly, Gary (2021) reframes ethical or moral critiques (e.g. the threat of diminished human interaction and social connectivity) within the socio-political context of capitalist societies. The critical issue, Gary (2021: 29) argues, is that capitalist enterprises rather than ‘collective self determination' are driving a radical transformation in the care of our ageing populations. What is at stake, ‘is not the inherent nature of humans but the capacity of individuals and communities to determine the direction of their lives' (ibid, 31). An ethics of care that centres ‘interdependence, attunement to the needs of others, attention to the particularities of context, and responsibility as grounded in relationships' (ibid, 31) is an alternative narrative that contradicts capitalist economic logic in (potentially) politically productive ways, although this narrative too can be co-opted to capitalist ends (Fraser, 2016). Conceiving care robots only as a response to the predicament of the liberal individual or a fundamental affront to what it means to be human attenuates this broader debate.

We turn first to introduce two companion robots, one developed in the USA and another by an Israel-based firm, both marketed in the first instance in the USA. They are two of the three companies identified in the 2022 Market Research Update report on the Robots for Seniors Market as the ‘key players of the market' (Market Insight Reports, 2022). We then discuss three robots developed by researchers/inventors in the UK (in the first two instances through substantial public funding), the challenges of bringing them to the market and the assumptions and aspirations that lie behind them. Finally, we ask whether and how we might reframe those aspirations.

Robots on the market

Joy for All

In 2016, the US-based toy manufacturer, Hasbro, noticed that about 10–20% of their robotic pet product, FurReal Friend, was being purchased for adults, not children. Marketed to ageing adults, Joy for All was launched within a new company called Ageless Innovation.⁴ They created a very simple robot (Figure 1). As an academic nurse/researcher put it: ‘So [it's] very, very low tech. You know, just a set of programmed movements: eye movements, ear movements, yawns, purrs, rolls over … I mean it reacts to sound and light, but isn’t programmable … So a low tech and furry cat. Really pretty little thing'.

Figure 1.
Joy for All pet. Photograph by authors.

Jim Murphy, Senior Vice President, Healthcare of Ageless Innovation, identifies two target consumers for their cat and dog robotic (and now bird, the Walker Squawker) pets, ‘one more obvious than the other'. ‘So, the one that was early and obvious was providing the companion pet to an older individual with Alzheimer's or related dementia and the positive impact that could be made there', notes Murphy, ‘And then the other one, which wasn’t so obvious – the one that had me really curious – was around “What about individuals without cognitive issues but who are still socially isolated or lonely?”' When Murphy joined the company, it was ‘really to add fuel', he says, to what he identified as ‘a foundational start [by Ageless Innovation] in the healthcare business primarily focused on … long term care, memory care, nursing homes, assistive living, Medicare Advantage’. The fuel caught fire: ageing departments in 21 states in the USA have now distributed more than 20,000 Joy for All pets. As of May 2022, the state government of Florida alone has purchased approximately 8500 pets (Engelhart, 2021). In 2020, an insurance company in Minnesota received approval to fund Joy for All pets for older policyholders, and the company is hopeful that private Medicare plans will follow suit. The journalist, Katie Engelhart, was told by one company executive that this is ‘everybody's holy grail'. In March 2022, the company began efforts to establish a formal relationship with the National Health Service in the UK. Ted Fisher, the CEO of Ageless Innovation, is quoted as saying, ‘If you live in Minnesota and fit into the categories of experiencing [loneliness or dementia …] it's a fully-covered supplemental benefit, and our goal would be to follow that in the UK’ (Williams, 2021). Joy for All pets are now sold in 30 countries, and the company reports annual revenues of $7 million, with triple-digit annual growth since 2018 (Halpert, 2021).

The success of Joy for All pets has occurred in relation to the anticipated costs of social isolation, calculated in an influential 2018 study by the American Association of Retired Persons (AARP), which estimated that social isolation in the USA costs $6.7 billion a year. In a presentation at the 2021 annual meeting of the American Society on Aging, Greg Olson, the then Acting Director of the Office of Aging in New York State, weighed those numbers against the expense of a Joy for All pet, which cost roughly $100 each. His decision to purchase 2200 robotic pets to distribute during the pandemic was also informed by his agency's own research. Deploying the De Jong Loneliness Scale at 3 months, 6 months and 12 months to measure outcomes, he reported a 70% reduction in loneliness and 75% reduction in pain among those in possession of a robotic pet. He stated that, in his 29 years of government service, the distribution of robotic pets is ‘one of my greatest projects'.

The role of research and university researchers in this development network has been extensive and decisive (see Bradford, 2021; Bradwell et al., 2019; Fogelson et al., 2022; Koh et al., 2022; Yu et al., 2022). An executive of Ageless Innovation says:
It's been absolutely critical … Most decision makers are looking for a quantitative evidence-based approach and so the research has been absolutely critical … You’ll see across the studies of psycho-social benefits and reducing loneliness and improving mood using a number of different measures … But then within two or three of the smaller pieces of research we’ll find that, for example, within facilities they noted that there was either a) a reduction in the need to administer psychotropic medications … (which is huge, but we’ve only proven it in small numbers. So, if we could prove that in larger numbers, then there could be a lot more adoption). And then, b) is falls-prevention. There's [documentation of] a reduction of falls in a couple of the studies. They were once again relatively small studies. You know, we need to continue to demonstrate value to be able to really get beyond what is an initial starting point.
The work of university researchers and research in networks of development and market expansion across a range of technologies is one – possibly less obvious – facet of the public funding of private enterprise.

Rather than replace human-to-human communication and social relatedness, the argument is made that Joy for All pets function to enable or to facilitate social connection. A nurse researcher in the UK described giving her mother a cat: ‘She was just looking at this cat as if somebody had given her a million pounds … And the carers started interacting with the cat as well, which was another thing and they started talking to the cat as they were talking to mum, and dad used to talk to cat and talk to mum. So, it became part of the family and part of the dynamic'. This description of a robotic companion animal meditating, easing and enhancing human interaction between those living with dementia and their caregivers, rather than replacing human interaction or the need for caregivers, recurred in other interviews and seems to come close to the kind of interaction that Atanasoski and Vora (2019) characterise as a social understanding of self and other, in which the content of communication emerges through bodies in relation to others (rather from within and between interiorised contained selves). This may very well be facilitated – in part – by design. Animatronic dogs and cats are not designed to possess or communicate complex emotions. The cats (more popular than the dogs possibly precisely because less emotional connection is expected) offer a pre-programmed purr and some silky fur to stroke. In this, they may offer some comfort in themselves and facilitate human interaction.

ElliQ

More is expected of ElliQ. Launched commercially in March 2022, ElliQ is a table-top stationary robot (with head that swivels as part of its non-verbal communication, for instance, bowing to communicate an apology) designed by the Israel-based company, Intuition Robotics. This robot is the CEO's fifth venture. He worked previously on projects in cybersecurity and ‘consumer stuff' and wanted to create something that would have a high social impact: ‘9 out of 10 start-ups fail so you might as well try to do something important'. It was clear to him, as well, that there was considerable room to innovate within the ‘longevity' market: ‘the issues are humongous and have almost no innovation around them'. It is ‘a wide space that could give the company a lot of maneuvering room'. He was also struck by an exercise that he goes through whenever he starts a company. He approaches experts to ask them a single question: ‘What is the biggest unsolved problem in your space?' He typically hears a variety of answers but in this case ‘something really interesting happened': 19 out of 20 interviewees said the same thing: loneliness and social isolation.

ElliQ was built to address this problem. We were told by the CEO that the development of ElliQ over the last 6 years has cost ‘10s of millions' of dollars. ElliQ ‘gives the end user the feeling that empathy is projected towards them' and answers the human social need ‘to be seen'. ‘She’ (the pronoun used by the CEO) is designed to be proactive, initiate conversations, take an active interest, read body language, engage the user with humour and motivate the user toward a more active and/or comfortable life. ElliQ typically greets her human companion in the morning, inquires about their sleep, may recommend appropriate exercise or other activity, takes an interest in and follows up on previous conversations, enquires about visits with friends and helps her companion to reach their goals. ElliQ's option menu is pre-programmed, but artificial intelligence (AI) decides on the response based on learning and the level of familiarity with the person. ElliQ has been purposely designed and built for the American older adult: the humour, historical references, interesting facts and play of words are all developed for this market and subject. The CEO quipped: ‘We can make her politer and bring her to Canada as well’. ElliQ is marketed primarily to people living alone. The robot is not built to be in shared spaces:
ElliQ is your friend. She gets to know you. She is not a gadget or a party trick. It's a one-on-one relationship … People really build a relationship with her. And it stays over time. It's not a problem to create something that people use and enjoy for a week or two … If you look at a single cohort of users, six months later they are still using ElliQ about 17 times a day.

We asked about the capacity of humans to have a meaningful relationship with ElliQ: ‘Clearly we can. We are wired to anthropomorphise and give meaning to things that don’t … As humans we just need a nudge'. But the designers have tried to make it clear that it is a relationship with a machine through its design, name and scripting for when there are ‘opportunities for confusion'. The CEO gave an example of the latter:
If somebody says ‘ElliQ, I love you’. What's the right answer? What is the right answer? So, the way we answer it is, first they give you some kind of recognition (there is a reason you said that): ‘Oh, thank you’. And then remind the person that she is not real. They’ll say something like: ‘That makes my processor overheat’. Or ‘Stop it, my fan will spin out of control’. We are a little bit light but reminding you that ‘We are both agreeing here to play this game, because it gives you comfort. But I am not real’.

The CEO of Intuition Robotics is clear that human care and ‘human touch' are preferable to care by robots, ‘when it is caring and not just a body that is paid to give you a function. But the fact of the matter is: in modern society we leave our elders alone. A lot'. He notes that even in the situation where caring children invite their parents to live with them, they likely work outside the home, and that parent will be alone 10 h a day. In this scenario, ElliQ is a good supplemental companion: ‘And that's an extreme good example. In most cases Mom is just alone. Full stop … I think it's inevitable, unless we go back in time, and live in a village, and treat our elders as the village wise man or woman, as opposed to a burden … Society in the last 200 years is moving in a very different direction. So, yeah, technology can help'. With a $249.99 one-time enrolment fee and $29.99 monthly subscription, ElliQ is there to help. As of June 2022, New York State Office for the Aging has purchased 800 ElliQ robots. The New York State Office for the Aging Director, Greg Olson, also instrumental in introducing Joy for All robots (see above), has said: ‘The goal is to measure. We’ll see what [ElliQ] does, and how it does' (Osborne, 2022). The CEO of Intuition Robotics is keenly aware of the importance of this research for marketing ElliQ, and the company is currently working with researchers at Duke University to create a companion to the UCLA Loneliness Scale: ‘like a digital companion efficacy score'.

Developing care robots

We turn to introduce three companies working on companion robots in the UK, the first explicitly addressing a present limitation of ElliQ. Intuition Robotics is for the moment focused exclusively on the US market, and the CEO is fully aware of the challenges of building an empathetic robot for other cultures, given the specificity of natural language processing, cultural and historical references and body gestures (e.g. the meaning of ElliQ's implied gestures, such as bowing), among other factors. The Japanese consumer market is compelling, given the large ageing population and generally favourable attitudes to robots. But Intuition Robotics currently weighs this against the significant development costs. The CARESSES project took cultural variation as its central focus.

CARESSES

Building cultural competency was the explicit task of Dr Irena Papadopoulas, a UK-based health researcher specialising in transcultural nursing and health, who worked within a consortium of six European Union (EU) partners and three Japanese partners on the CARESSES project, which was funded (€2,084,248) by Horizon 2020, the EU's research and innovation funding programme, and the Japanese Ministry of Internal Affairs and Communications (¥60,000,000 or €443,400) from 2017 to 2020. The fact sheet for CARESSES states that the research was the ‘first ever attempt to develop and test a culturally-aware robot' and will ‘expand the capabilities of the Pepper robot, which was designed and marketed by SoftBank Robotics, a partner of the project'. (SoftBank Robotics Europe was one of the six EU partners.) In 2018, Pepper became the first non-human witness to give evidence live to the UK Parliament on how CARESSES could help ease the burden of elder care in the UK.⁵

The fact sheet for the CARESSES project rehearses the common assumptions that drive robot development: ‘Ageing populations across the world are placing health systems under increasing pressure. Elderly care robots can be a means to relieve that pressure in hospitals and care homes, as well as a way to improve care delivery at home and promote independent living for the elderly'. Building cultural competence will likely create ‘greater acceptance [of robotic technology] from both the elderly and their caregivers' (CARESSES Factsheet).

Building this greater acceptance is important to SoftBank. Introducing Pepper in 2014, SoftBank made claims equivalent to those made for ElliQ. According to SoftBank (2016), ‘Pepper's not here to replace humans. Or even vacuum the floor. Pepper is here to make people happy, to help them grow and enhance their lives'. Pepper can remind you to take out the garbage, take your medication, engage you in conversation or give you a hug. SoftBank advertises Pepper as ‘friendly and engaging. Pepper creates unique experiences and forms real relationships'.⁶ Nevertheless, Pepper has not been a market success, likely due to its high cost (roughly $20,000), along with limited functionality. Only 27,000 Peppers were produced during the 6 years of production, and SoftBank, struggling to find a customer base, halted the production of Pepper in 2020, with promises to restart production when excess supply is sold (Nussey, 2021). A researcher investigating human–robot interactions within an Adaptive Systems research group said: ‘Pepper might look nice in the short-term experiment, but we also found out that this is caused by … the novelty effect … It's totally looking nice. It's a funny thing. It talks to you. It's smiling. These kinds of things. It's interactive. But then [experiment participants] come for a second and third time and it becomes normal [and less interesting]. And if it's in your house it's similar'. The anticipated emotional connection, in other words, was not found to happen. A woman working for a non-profit providing services for persons living with dementia in the Northeast of England remembered watching a video of Pepper:
It showed them doing household tasks. Very slowly. [Laughs.] And it did show some interactions, some conversations. So, it might have sat quietly not functioning but if the person spoke it would attempt to make a conversation … We spent a long time watching them and I think I sat there wanting to say, ‘Hang on a minute, have you … spoken to anybody yet who's actually got dementia and knows what it feels like?’ And I think I probably sat there feeling a bit angry, so I didn’t take much of [the video] in.

The CARESSES project addressed this gap insofar as extensive field testing was built into the project. Dr Papadopoulas’ role was to find ways ‘that the robot would understand the cultural background, the cultural identity of the older people that they were with'. She wrote and tested cultural scripts for three groups: British English, British Hindu and Japanese. She began by writing stories from everyday life that captured ‘all the values, the beliefs, the attitudes, all the things that you use, even though they may not be conscious of them, but when the time comes, you use them to make decisions on how to conduct your life'. From these stories, they wrote detailed scenarios: ‘We had to write about every detail of what the human does and what the robot should do … There's pages and pages of this which eventually went into the software, into the programming of the robot'. They video-recorded older people in participating UK and Japanese care homes interacting with Pepper at six different times of the day and also collected data on ‘enactments’ of emotions.

The research team validated the scripts by creating short videos from this extensive archive of video footage for each of the three cultural groups. They showed these short videos to new samples of older people, asking them to fill out a short survey inquiring if the behaviours represented their cultural group. They developed a model to enable the robot to more effectively learn from interactions with older people. Dr Papadopoulas explained:
ADORE stands for Assess, Do, Observe, Review, Evaluate. These were the steps that the robot would do. The robot would [for example] go near the person and say ‘Oh, I see you’re, I feel you’re looking a bit sad today, Mrs. Smith. Would you like a cup of tea?’ The robot knows what kind of questions to ask for assessments. But it selects them by him, by itself. Totally autonomous. So, Mrs. Smith says, ‘No, I’m not sad. I’m just tired. I’ve just had a cup of tea. Thank you’. So, the robot says, ‘Well, would you like me to bring you a glass of water then maybe’. And she says, ‘Oh, that will be nice’ … In the guidelines, we have a lot of different examples of using the ADORE. The robot [also] has other indicators for being culturally competent, for example, culturally appropriate volume of voice and conversational distances.
Dr Papadopoulas spoke of the robot's capacity to address loneliness:
So, the robot at the end was very capable of holding a conversation. And what the robot did well was to be a good companion to the older person who was lonely, or maybe not lonely but still enjoyed having a talk … We’ve interviewed people in care homes as well and they … feel like they’ve been abandoned even though the care is good. They’re missing their home. And they’re missing seeing the family more frequently. Because the family now thinks they’re safe and they’re cared for, and so they don’t visit so often. So [the robot is] very good with that.

The robot can also help by offering a range of activities and motivating the user (‘Come on Mrs. Smith, you can do that, jump higher') and by connecting the user to family and services. Dr Papadopoulas’ impression is that the residents at the care facility where they conducted their research enjoyed the companionship during the field trial: ‘A lot of the older people who had the robot for two weeks actually missed it after it was no longer there. So, they had to do some debriefing'. The field trial established that those with the robots were happier and felt less lonely: ‘Because loneliness is really the issue'.

The Advinia HealthCare care homes, where the UK field trials were held, ‘have been dying to have to have Peppers all over their homes. But unfortunately, when we finished the prototype, we didn’t have any money to take it to the market. In order to actually commercialize it, there are major processes, and you need big money, and you need investors. So, we got a lot of awards and a lot of praise by the funders, and we had high hopes. But anyway … We need the infrastructure. Governments have to … We want the robots, we want to have all this technology in our workplaces and all of that. But they’re just words. You have to create the infrastructure'. This disconnect between massive public investments in robot development is common: the Director-General of the Research Institute of Science and Technology for Society (RISTEX) in Japan refers to this as ‘the “valley of death” between engineering and social impact' (quoted in Wright, 2021: 9). The EU and Japanese governments have invested hundreds of millions of euros (or yen) in the last 20 years (in the case of the EU, approximately €2 billion over the last decade under Horizon 2020, the funding source for CARESSES, and the Active and Assisted Living Programme (AAL)), with a similar disconnect between what Wright (2021) identifies as ‘top-down' approaches to funding development and ‘bottom-up' approaches to implementation. The Japanese government's strategy has shifted in recent years, from funding robot development through university-based research consortia to setting specifications in relation to perceived needs of the care industry and funding private companies directly for robotic development: ‘The prioritization of treating care robotics as an industry rather than as a research domain may help explain the greater number of robotic care products commercialized through publicly funded Japanese research projects' (Wright, 2021: 11). With SoftBank incorporated as a research partner in the CARESSES project, the lines may be blurry in the EU as well. Both scenarios – funding through universities or directly to industry – raise questions about the public financing of private commercialisation.

Dr Papadopoulas also notes the limitations of robots for the care of ageing people:
These don’t solve the issue of older people. Older people need not only companion[ship] and health promotion and all of that. They need care. The issue about physical care, washing the older people, feeding the older people …, taking them to the toilet, these [robots] don’t exist. And I can’t see how they’re going to exist for a long time.

In an ethnography of the trialling of Pepper in a public care home in Japan, James Wright found this to be the case. Indeed, Pepper was effective in leading exercise and other recreational classes, but ‘it was not realistic that Pepper could act in a standalone way, roaming the halls and interacting with residents … Like residents themselves, Pepper was dependent on the care of the staff' (Wright, 2019: 346). Perhaps, the most interesting issue is that the capacity for robots to act as effective companions and motivators is not in doubt; it is the less skilled, more physically taxing and degraded tasks (such as toileting) for which humans are currently irreplaceable, replaying in a novel way Atanasoski and Vora's (2019) suspicion that automation often upholds and extends historically sedimented structures of racial, gender and class hierarchy and inequality.

Robbie

Beyond social companionship, we spoke to designers creating predictive companion technology. Robbie is a humanoid robot designed at Edge Hill University in Lancaster, trained to spot signs of depression and anxiety in persons living with dementia. The creator, Dr Ardhendu Behera, previously developed AI to detect abnormal behaviour (and hence security risks) at baggage claims in airports and came to this project obliquely by attending a Health Network meeting that gathered university researchers together with healthcare organisations to ‘brainstorm how technology can help in the healthcare domain'. Like so many others, Dr Behera frames his current work within the costs of dementia care: ‘Currently the only ways to monitor and manage dementia is by direct observation, which is labour intensive, time consuming and can be costly from a care perspective. Or there's wearable bio-sensing devices. Monitoring and recognition are still very much in their infancy and we believe Robbie is the first robot to use vision-based recognition to recognise four behaviours: aggressive, depressive, happy and neutral'. Thus far, Robbie has learned to detect these behaviours in persons living with dementia by watching 13 episodes of the British television soap opera programme, Emmerdale, in which one of the main characters has vascular dementia (Figure 2). Dr Behera noted: ‘The Alzheimer's Society in UK formally appreciated the acting. Ashley Thomas's acting is quite an accurate replication of what a person with vascular dementia would behave like: that aggressive behaviour. We took those 13 episodes, and trained our model based on his normal behaviour and aggressive behaviour'. (On the Alzheimer's Society's reactions to the Emmerdale portrayal of dementia, see BBC News (2016); see also Morris and Morris (2021) for the programme's positive portrayal.) As for testing the robot in a real-life setting, Dr Behera noted the ethical challenges. Added to this is the practical issue of the ‘bottom-up’ approach to the implementation typical of companion robot development. As Wright (2021: 8) notes in relation to EU-funded projects (of which, Wright claims, all UK projects were before Brexit): ‘decisions about trialling, leasing or buying … tend to be local government commissioners, care home managers, care managers or older adult users themselves and their relatives or care providers'.

Figure 2.
Robbie learning about dementia. Photograph by Edge Hill University.

In an effort to reach a market, Dr Behera has partnered with Home Instead Care, a large US-based homecare company, and he envisions training Robbie to detect a wider set of behaviours and pending care problems: ‘For example, if you are sitting on a chair for a long time, and there is no movement at all. Okay, so then Robbie might go closer to look at whether you are really sleeping, or something is not right with you. So, these are the types of scenarios we look at. [… Or] if you fall down, Robbie will call the loved one of the one who has fallen down on the floor. [He will] dial their contact number and contact them to help … But to predict something in the future, based on the work we are doing currently, is a very difficult challenge for most researchers in artificial intelligence and data science. So that's the kind of future we are looking for'. Dr Behera's partnership with Home Instead Care is with the company's technology experts, who specify the company's plan for introducing technology into domicile care: ‘I wouldn’t say that I have much information from the home perspective. It is more from the technological perspective; of what feasibility they are looking for'. This replays the norm of designing technology for – rather than with – the users (Berridge et al., 2022; Morrissey et al., 2017). It replays, as well, the scenario of the isolated ageing person needing care in their private home.

ALTAIR

We spoke with the inventor of the third companion robot prototype. Tony Ellis built one of the first programmable robots in 1979, aptly called ‘Vintage’. Now retired, one of his most recent humanoid robot prototypes is called ALTAIR (Figure 3). ALTAIR can hold a conversation and express 26 emotions through expressions of mouth and eyes and detect emotions through facial recognition software: ‘ALTAIR's got its own emotion engine. You know, it gets sad if you’re sad, and so on'. In a care situation at night, ‘it would automatically go back to the person's bedroom, where it would be on its charger. But even though it would be in a shutdown mode, its sensors would still be listening for problems. Like if it heard a shout, or if there was a … Because it's got thermal imaging, it could see if the person was in pain. And I’m testing micro radar down to a point where, without contact with the person, it can tell that they’re breathing properly, … can see if you move off the bed, or fall off the bed, and so on. So, the robot is like, a companion. That's what I exactly what I wanted it to be, a companion for the elderly'.

Figure 3.
ALTAIR's emotional responses. Photograph by Tony Ellis.

In Tony Ellis's view, the one job that cannot be taken over by a robot is moving a patient: ‘You can’t touch moving a patient because it's uninsurable. The litigation'. What a robot can do well is telemedicine and monitoring health in the home (by monitoring without touching, for instance, an individual's face, blood pressure, blood oxygen and temperature): ‘And the other big thing is companionship. Companionship is something that's really missing, especially if you’re an elderly person alone'. ALTAIR has a conversation engine that allows it ‘to make conversations that are fresh, and not scripted. Over time, it learns what the person likes. Say Rose loves EastEnders, the TV show. What happens is when you pair up a robot with the user, you just put in a big list of their likes, and it starts this whole programmed algorithm. For instance, if the robot is looking at RSS feeds maybe it would see some information about Dot Cotton [an EastEnders character]. What the algorithms do, they … build it into a conversation. So, the robot might come back and say, “Oh, did you know that Dot Cotton was leaving the show?” You know, it's stuff like that. So, it's not great conversation, but it's not all fixed. So, you don’t get the same thing all the time and you get some new data into the conversation. Conversation is a huge thing in companionship and loneliness, you see … Another area that I want to work on is bad news, good news. What I don’t want is the robot to give bad news. That's really difficult: to parse bad news and good news. But there are algorithms that can do it. I mean, the last thing you want the robot to say is, “Oh, a plane's crashed and 57 people are dead.” You don’t want that. You want just positive obviously'. In Tony Ellis's view: ‘We’re not going to ever replace carers. No way. Can never happen. But if it can help or replace two four-hour shifts a day. At what the carers get, which is the universal wage limit – which is £8.00 or something like that – the ALTAIR could be paid back in less than six months'.

As is the case more generally, Tony Ellis's imagination was fuelled by popular culture: ‘ALTAIR is the dream I’ve had since I was a child. You know, when I first saw Robby the Robot from Forbidden Planet … that was one of my favourite films'. (First appearing in this 1956 film, with subsequent appearances in other films and televisions programmes, Robby has been dubbed ‘the hardest working robot in Hollywood' (Telotte, 2016).) Musing about the possibility of having a genuine caring relationship with a robot, Tony Ellis took us back to popular culture: ‘I think yes [a genuine caring relation with a robot is possible]. Do you know what, there's some lovely videos [and films] that have been on recently. Robot and Frank … Bicentennial Man, that's another, with Robin Williams … Again, that's another one where robots are taken into the family. In fact, in that case, Robin Williams – Bicentennial Man – becomes a member of the family. And Tim Spall has just done this wonderful short film [This Time Away]. He's given a robot, and the robot basically starts looking after him. You’re in tears at the end of these things, because, I mean, they are very emotionally powerful’.

Imagining futures of care

How should we think about this future of chatty, joke-cracking, current affairs-savvy companion robots, with humans labouring at £8.00 an hour in the physically taxing jobs of lifting, toileting and feeding the bodies of ageing selves? In a strange reversal of humans and machines, robots seem designed to do the cognitive and emotional care work, humans the manual work of bodily care.⁷ It is clear that technology has been and will be designed to offer a kind of companionship and care. Considerable industry and academic research verifies that companion robots – even a low-tech fluffy cat – offer comfort and an improved sense of well-being. This has been critical to the marketing and publicly funded uptake of companion robots. There is a sense that these technologies work best in relation to human care and that more relational interdependent human–machine relationships are possible (Atanasoski and Vora, 2019; Dobson, 2007; Lynch et al., 2022; Sumartojo and Lugli, 2021; Woods and Kong, 2020). There is comfort imagining the Joy for All cat or ElliQ in a relationship with rather than replacing a human caregiver. But as the CEO of Intuition Robotics (creators of ElliQ) reminds us: this is the best-case scenario.

Much like the imaginative constructions that have carried robot designers along to create the likes of ALTAIR, what cultural assumptions are carrying us along to create a future where robots seem like an attractive or inevitable emotional care choice for ageing adults? If we have the imagination to envision and create a future of extra-human care, in what other ways do we need to stretch our imaginations in order to think into being the social worlds we want to inhabit, ones in which ElliQ is a supplement to care rather than the primary life companion?

Loneliness is taken as a given in much of the discussion of companion robots. There is a small scholarly industry of ‘scoping reviews’ on care robots for elderly persons, and these reviews invariably identify loneliness as a significant guiding rationale for the development and use of robots for the care of elderly persons (e.g. Agraz et al., 2022; Budak et al., 2021; Hung et al., 2022; Koh et al., 2021; Sather et al., 2021; Todd et al., 2022; Yu et al., 2022). In her biography of loneliness, historian Alberti (2019: vii) notes that we now find ourselves in the midst of a declared loneliness epidemic, necessitating the creation of the Minister of Loneliness in the UK in January 2018: ‘Talking about loneliness seems to spread, contagion-like, until it has become part of the social fabric. Certainly, it has become a convenient hook on which to hang a number of discontents' and, we might add, public funding of companion robots. All emotions are political, notes Alberti (2019: 229), ‘but at this historical moment, none is so political as loneliness', in part because it has been monetised in recent years and found to be extremely expensive. Loneliness (as distinct from aloneness) has a history, constructed since the 19th century, through modern scientific beliefs about the body and mind, an ideology of possessive individualism (MacPherson, 1962), a growing (gendered) division between public and private and the increased incidence of living alone (Snell, 2017). In the 19th century, Alberti (2019: 229) notes that ‘a new form of emotional experience was born'. With the spread of possessive/competitive individualism, ‘a vacuum had emerged in which the self was alone, marooned, and dependent on familial and social networks that were [also] in a state of flux' (Alberti, 2019: 229). Identification of old people as a distinct (pathologised) social group also has a history and geography (Alberti, 2019; Roebuck, 1979). It is a social category that does not do well within a neoliberal ethos that defines worth by one's economic contribution. One point of remembering these histories of loneliness and old age is to remind ourselves that the present could be otherwise and that our present conditions have arisen through myriad socio-political and economic choices that could be taken differently.

Some of the clearest choices pertain to the orientation of public monies and public policy. Decisive to the funding of robotic development is the fact that, as Wright (2021: 9) notes, it is more acceptable for conservative (or neoliberal) governments to invest in innovation than the day-to-day delivery of social care services. Alberti (2019: 233) observes that ‘at the same time that the [UK] government created a Minister of Loneliness, who proposed to engage communities and understand loneliness as a human experience, it continued to strip the assets and spaces where community was being formed, especially in the poorest sectors of society'. These included spaces, such as public libraries and council housing, and social care services. Funding innovation to replace social care services is the more acceptable option.⁸

Alongside the poverty of imaging community spaces as spaces of care is an idealisation of the individual home as the space in which ageing persons in need of care are best served, as well as the high value given to personal autonomy. When Reid (2022) interviewed representatives of Scottish local governments during the pandemic, she found that independence and empowerment were the key desired outcomes that they considered when planning technology-enabled homes; the importance of interdependency was rarely raised. We have lived within individualism and neoliberal valuations of the self-actualising, ‘responsibilized’ self for generations now, such that it is naturalised. But there have been and are other ways of living. As Alberti (2019: 155) notes, calling up the early work of the geographer Laws (1993) on ‘ageist’-built environments, segregating housing by age is a modern invention that ‘both pathologizes the status of being old and presupposes homogeneity between people on the basis of age'. Might we imagine something between this and ‘independence’ in a private home in the company of a robot? As a defence against social isolation, automation can forestall discussions of what is possible in less individualised, more congregate forms of housing or more supportive and publicly supported community arrangements, including a range of community services and spaces. Critiques of care robots can reopen these discussions.

Included in a reimagining of care of ageing persons is a critical revaluing of care more generally. It requires the expansive rereading of capitalism that has been called for by feminist political economists for generations, in which care work is written into a well-functioning economy rather than hived off as a separate non-valued (and hence undervalued) sphere. Can we imagine a world in which robots augment the labour of well-paid care workers rather than one in which the work of caring for our ageing population is so devalued that robots replace them and leave the worse aspect of the job to minimally paid racialised immigrant workers? This is not a hypothetical or rhetorical question: it is the question that confronts us in our current situation, if we choose not to ignore it.

Footnotes

Acknowledgements

We thank Donna Baines and anonymous reviewers for their constructive engagements with this paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada, as well as a Faculty Research Grant from Newcastle University.

Notes

References

Agraz

Pfingsthorn

Gliesche

, et al. (2022) A survey of robotic systems for nursing care. Frontiers in Robotics and AI 9: 1–18.

Alberti

(2019) A Biography of Loneliness: The History of an Emotion. Oxford: Oxford University Press.

Atanasoski

Vora

(2019) Surrogate Humanity: Race Robots, and the Politics of Technological Futures. Durham and London: Duke University Press.

Berridge et al. (2022) Control matters in elder care technology: Evidence and direction for designing it. In: Designing Interactive Systems Conference (DIS '22), June 13-17, Virtual Event, Australia, 26 pages.

BBC News (2016) Emmerdale dementia episode praised by Alzheimer’s Society, 21 December. https://www.bbc.com/news/entertainment-arts-38389953 (accessed 5 August 2022).

Berridge

Turner

Zaslavsky

, et al. (2022) Design for dementia: A call from the health sciences. Interactions 29(2): 47–49.

Bradford

(2021) Robot pets in dementia care: an educational intervention . Unpublished Doctor of Nursing Practice Project. University of Massachusetts Amherst.

Bradwell

Edwards

Winnington

, et al. (2019) Companion robots for older people: Importance of user-centred design demonstrated through observations and focus groups comparing preferences of older people and roboticists in South West England. BMJ Open 26 (9): e032468.

Budak

Atefi

Hoel

, et al. (2021) Can technology impact loneliness in dementia? A scoping review on the role of assistive technologies in delivering psychosocial interventions in long-term care. Disability and Rehabilitation: Assistive Technology: 1–13. DOI:10.1080/17483107.2021.1984594.

10.

CARESSES Factsheet. http://caressesrobot.org/en/wpcontent/uploads/sites/2/2017/10/Fact_sheet_v3.0.pdf

11.

Carros

Schwaninger

Preussner

, et al. (2002) Care workers making use of robots: Results of a three-month study on human-robot interaction within a care home CHI 2022 (Conference on Human Factors in Computing) April 29-May 5, New Orleans.

12.

Del Casino

Jr (2016) Social geographies II: Robots. Progress in Human Geography 40: 846–855.

13.

Dobson

(2007) Machine therapy . Unpublished PhD. School of Architecture and Planning, Massachusetts Institute of Technology.

14.

Engelhart

(2021) What robots can – and can’t – do for the old and lonely. The New Yorker. 31 May.

15.

Federici

(1975) Wages Against Housework. Bristol (UK): Falling Wall Press.

16.

Federici

(2012a) On elder care. The Commoner 15: 235–261.

17.

Federici

(2012b) The unfinished feminist revolution. The Commoner 15: 185–197.

18.

Fogelson

Rutledge

Zimbro

(2022) The impact of robotic companion pets on depression and loneliness for older adults with dementia during the COVID-19 pandemic. Journal of Holistic Medicine 40: 09.

19.

Folbre

(2001) The Invisible Heart: Economics and Family Values. New York: The New Press.

20.

Folbre

(2012) For Love and Money: Care Provision in the United States. New York: Russell Sage Foundation.

21.

Fraser

(2016) Contradictions of capital and care. New Left Review 100: 99–117.

22.

Fraser

(2022) Cannibal Capitalism: How Our System Is Devouring Democracy, Care, and the Planet – and What We Can Do About It. London and New York:Verso.

23.

Gary

(2021) Care robots, crises of capitalism, and the limits of human caring. IJFAB: International Journal of Feminist Approaches to Bioethics 14(1): 19–48.

24.

Halpert

(2021) How a Hasbro exec made millions selling robot pets to adults. Business Insider 30 September. https://www.businessinsider.com/how-hasbro-exec-started-business-selling-robotic-pets-adults-2021-9. (accessed 8 August 2022)

25.

Herbrechter

(2021) Before Humanity: Posthumanism and Ancestrality. Leiden and Boston: Brill.

26.

Hung

, et al. (2022) Facilitators and barriers to using telepresence robots in aged care settings: A scoping review. Journal of Rehabilitation and Assistive Technologies Engineering 9: 1–19.

27.

Kamphof

(2015) In the company of robots: Health care and the identity of people with dementia. In: Swinnen

Schweda

(eds) Popularizing Dementia: Public Expressions and Representations of Forgetfulness. Bielefeld: Transcript, 359–375.

28.

Koh

Ang

FXH

Casey

(2021) Impact of low-cost robotic pets for older adults and people with dementia: Scoping review. JMIR Rehabilitation and Assistive Technologies 8: e25340.

29.

Koh

Whelan

Heins

, et al. (2022) The usability and impact of a low-cost pet robot for older adults and people with dementia: Qualitative content analysis of user experiences and perceptions on consumer websites. JMR Aging 5: e29224.

30.

Laws

(1993) The land of old age’: Society’s changing attitudes toward urban environments for elderly people. Annals of the American Association of Geographers 83: 672–693.

31.

Leland

(2022) Can robots save nursing homes? The New York Times, 21 April. https://www.nytimes.com/2022/04/21/realestate/nursing-home-robots.html?searchResultPosition=5

32.

Lynch

Bissell

House-Peters

, et al. (2022) Robotics, affective displacement, and the automation of care. Annals of the American Association of Geographers 112: 684–691.

33.

Macpherson

(1962) The Political Theory of Possessive Individualism: Hobbes to Locke. Oxford: Oxford University Press.

34.

Market Insight Reports (2022) Robots for Seniors Market (US, Europe, Asia-Pacific), May.

35.

Meehan

Strauss

(eds) (2015) Precarious Worlds: Contested Geographies of Social Reproduction. Athens Georgia: University of Georgia Press.

36.

Mezzadri

Newman

Stevano

(2022) Feminist global political economies of work and social reproduction. Review of International Political Economy 29(6): 1783–1803.

37.

Mitchell

Marston

Katz

(eds.) (2004) Life’s Work: Geographies of Social Reproduction. Malden, MA: Wiley Blackwell.

38.

Morris

(2021) The Family Experience of Dementia: A Reflective Workbook for Professionals. London and Philadelphia: Jessica Kingsley Publishers.

39.

Morrissey

McCarthy

Pantidi

(2017) The value of experience-centred design approaches in dementia research contexts. In: Proc. of the 2017 CHI Conference on Human Factors in Computing Systems, 1326–1338. New York: ACM. DOI: 10.1145/3025453.3025527

40.

Nussey

(2021) SoftBank shrinks robotics business, stops Pepper production. Reuters 29 June. Available at: https://www.reuters.com/technology/exclusive-softbank-shrinks-robotics-business-stops-pepper-production-sources-2021-06-28/ (accessed 7 August 2022).

41.

Osborne

(2022) New York State purchases robot companions for the elderly. Smithsonian Magazine, 22 June. (accessed August 8 2022).

42.

Reid

(2022) Anticipating technology-enabled care at home. Transactions of the Institute of British Geographers 47: 108–122.

43.

Roebuck

(1979) When does old age begin? The evolution of the English definition. Journal of Social History 12: 416–428.

44.

Sather

IIIR

Soufineyestani

Imtiaz

, et al. (2021) Assistive robots designed for elderly care and caregivers. International Journal of Robotics and Control 3: 1–12.

45.

Schwiter

Steiner

(2020) Geographies of care work: The commodification of care, digital care futures and alternative caring visions. Geography Compass 14: n.p.

46.

Sharkey

(2012) Granny and the robots: Ethical issues in robot care for the elderly. Ethics and Information Technology 14: 27–40.

47.

Skills for Care (2022) The size and structure of the adult social care sector and workforce. Available at: https://www.skillsforcare.org.uk/adult-social-care-workforce-data/Workforce-intelligence/publications/national-information/The-size-and-structure-of-the-adult-social-care-sector-and-workforce-in-England.aspx (accessed 10 August 2022)

48.

Snell

KDM

(2017) The rise of living alone and loneliness in history. Social History 42: 2–28.

49.

SoftBank (2016) Available at: https://www.youtube.com/watch?v=oDeQCIkrLvc

50.

Sparrow

(2006) In the hands of machines? The future of aged care. Minds and Machines 16: 141–161.

51.

Sumartojo

Lugli

(2021) Lively robots: Robotic technologies in COVID-19. Social and Cultural Geography: 23 (9): 1220–1237.

52.

Telotte

(2016) Robby the robot and robotic persistence. Mosaic: An Interdisciplinary Critical Journal 49: 19–37.

53.

Todd

Bidstrup

Mutch

(2022) Using information and communication technology learnings to alleviate social isolation for older people during periods of mandated isolation: A review. Australasian Journal of Ageing 41: n.p.

54.

Turkle

(2017) Empathy machines: forgetting the body. In: Tsolas

Anzieu-Premmereur

(eds) A Psychoanalytic Exploration of the Body in Today’s World. London: Routledge, 17–27.

55.

Sommerlad

Sakure

, et al. (2022) Socially assistive robots for people with dementia: Systematic review and meta-analysis of feasibility, acceptability and the effect on cognition, neuropsychiatric symptoms and quality of life. Ageing Research Reviews 78: 101633.

56.

Williams

(2021) Robotic pets company hopes to work with NHSS to help older people living with dementia. inews, 15 February. Available at: https://inews.co.uk/news/technology/robotic-pets-company-hopes-work-with-nhs-help-older-people-living-dementia-806729 (accessed 7 August 2022).

57.

Woods

Kong

(2020) New cultures of care? The spatio-temporal modalities of home-based smart eldercare technologies in Singapore. Social & Cultural Geography 21: 1307–1327.

58.

Wright

(2018) Tactile care, mechanical hugs: Japanese caregivers and robotic lifting devices. Asian Anthropology 17: 24–39.

59.

Wright

(2019) Robots vs migrants? Reconfiguring the futures of Japanese institutional eldercare. Critical Asian Studies 51(3): 331–354.

60.

Wright

(2021) Comparing public funding approaches to the development and commercialization of care robots in the European Union and Japan. Innovation: The European Journal of Social Science Research: 1–16. DOI: 10.1080/13511610.2021.1909460.