AI & Soc (2016) 31:445–454

DOI 10.1007/s00146-015-0625-4

ORIGINAL ARTICLE

Robots in aged care: a dystopian future?

Robert Sparrow1

Received: 21 June 2015 / Accepted: 6 October 2015 / Published online: 7 November 2015
 Springer-Verlag London 2015

Abstract In this paper I describe a future in which per- 1 Introduction

sons in advanced old age are cared for entirely by robots
and suggest that this would be a dystopia, which we would Robot butlers have long been a staple of science fiction.
be well advised to avoid if we can. Paying attention to the The idea that we could rely on a robot assistant to help us
objective elements of welfare rather than to people’s hap- with daily tasks as we get older and frailer and thus help us
piness reveals the central importance of respect and stay in our own homes longer is an attractive one. It is this
recognition, which robots cannot provide, to the practice of thought—or, perhaps, fantasy—that appears to motivate
aged care. A realistic appreciation of the current economics many robotics researchers and the government agencies
of the aged care sector suggests that the introduction of that fund their attempts to develop robots for aged care.1
robots into an aged care setting will most likely threaten One way of responding critically to this project is to
rather than enhance these goods. I argue that, as a result, note just how far the technology is from rendering it
robotics for aged care is likely to transform aged care in plausible. A realistic understanding of what robots can and
accordance with a trajectory that leads towards this dys- can’t do—and what they are likely to be able to do in the
topian future even when this is not the intention of the future—is essential to any assessment of the prospects for
engineers involved. While an argument can be made for the them playing a useful role in meeting the needs of citizens
use of robots in aged care where the people being cared for in advanced old age (Sorell and Draper 2014; Sparrow and
have chosen to allow robots in this role, I suggest that Sparrow 2006). However, it is always possible that any
overemphasising this possibility risks rendering it a self- particular pessimistic estimation of the likely future
fulfilling prophecy, depriving those being cared for of capacities of robots will be overtaken by progress in sci-
valuable social recognition, and failing to provide respect ence and engineering. Perhaps more importantly, there is a
for older persons by allowing the options available to them risk that concentrating on the technical questions about the
to be shaped by the design choices of others. likely capacities of robots will blind us to deeper problems
with this project that would persist even if robots could do
Keywords Ethics
Robots
Robotics
Aged care
everything enthusiasts hope for them. Another line of
Society
Social robotics investigation, then, would be to consider what it would be

1
Notable contemporary examples include the European Union-
funded ACCOMPANY Project, which aims to build an ACCOM-
PANY System (or Care-O-Bot 3), which, according to the list of
objectives on its website, will provide ‘‘services to elderly users in a
& Robert Sparrow motivating and socially acceptable manner to facilitate independent
Robert.Sparrow@monash.edu living at home’’ [See http://accompanyproject.eu/ (accessed 16.1.15)]
and the (also EU funded) HOBBIT project, the goal of which is ‘‘to
1
Department of Philosophy, School of Philosophical, advance towards a robot solution that will enhance wellness and
Historical and International Studies, and ARC Centre of quality of life for seniors, and enhance their ability to live indepen-
Excellence for Electromaterials Science, Faculty of Arts, dently for longer at their homes.’’ [See http://hobbit.acin.tuwien.ac.at/
Monash University, Clayton, VIC 3800, Australia index.html (accessed 21.1.15)].

123
446 AI & Soc (2016) 31:445–454

like were engineers to be successful in building aged care the midst of all this robotic bustle, two robots in
robots. particular stand out: the telemedicine robot, which
I would therefore like to begin this paper by telling you allows medical personnel situated in a call centre in
a story. India to diagnose conditions, prescribe and adminis-
ter medications, and perform simple operations; and,
the telepresence robot, which allows relatives to talk
2 A visit to a robotic nursing home with and ‘‘visit’’ their parents and grandparents
without leaving the comfort of their own homes.
Imagine that you are visiting my university, Monash One might expect that this building would be silent or
University, for the first time. You are in a taxi, disturbed only by the buzzing of the robotic vacuum
travelling through the light-industrial area in which cleaners. In fact, it is filled with conversation and
Monash is located when you notice a long white laughter as the residents talk to their robot compan-
building sandwiched between two factories. There ions, which have been programmed to entertain and
are no windows on this building and from the outside converse with them in a never ending, if sometimes
it is hard to tell whether it is a warehouse, a factory, repetitive, stream of conversational gambits and
or a factory farm—although the cluster of antenna chitchat. The residents—especially those whose
sprouting from the roof suggest that whatever it is, it medical records show they have dementia—seem
involves the transmission of large amounts of data. happy. So effective are this facility’s operations
Careful observation would reveal that this building is that—apart from those it ‘‘cares’’ for—you are the
visited daily by several trucks and small vans; the first person to set foot in it for five years.
absence of any windows in these vehicles gives away
This story is science fiction.2 Indeed, for reasons I will
the fact that these are autonomous vehicles, the
discuss further below, it is more far-fetched than much of
commercial descendants of ‘‘Google car’’.
the reporting of current research on robotics, which is filled
You are curious enough to stop the taxi and get out
with glowing portrayals of the achievements and potential
and approach the building, the doors of which open
of robots for aged care, might suggest. Nevertheless, it is a
silently as you do so. Stepping inside, you realise that
recognisable extension of the sorts of claims commonly
it is an aged care facility for individuals with limited
made in the literature about the prospects for companion
mobility. There are no windows because each resi-
robots and/or service robots in aged care.3 Indeed, I hope
dent’s room features a number of window-sized
you will recognise many of the technologies I have inclu-
televisions displaying, for the most part, scenes from
ded in this scenario from the other contributions to this
some of the most spectacular parks and gardens
special issue; it is a world in which, I want to suggest, the
around the world. You do notice, however, that sev-
engineers have ‘‘succeeded’’.
eral residents appear to have set these screen so that
I have begun with this vignette for four reasons.
they show what they would have seen if they did have
First, although it is science fiction, I am also convinced
windows.
that it is dystopian science fiction: it describes a situation
What is most striking about the facility, though, is
that we should try to avoid rather than one to which we
that apart from the residents there is no one there. The
should aspire. Moreover, as I will argue further below, this
building is fully automated, staffed only by robots.
may remain true even if residents cared for by robots are
Robot sweepers, polishers, and vacuum cleaners
happier than they would be if they were cared for by human
clean the floors. Residents are turned and lifted out of
beings.
bed by the beds themselves, which can perform these
Second, I want to explore why this is the case. I will
actions either as a result of voice prompts from the
suggest that paying attention to the objective elements of
resident, remote instructions, or pre-programmed
welfare rather than to people’s happiness reveals the cen-
schedules. Sophisticated wheelchairs with autono-
tral importance of respect and recognition to the practice of
mous navigation capabilities move the residents
aged care and that the introduction of robots into an aged
around the facility, to the dining hall where pre-
care setting will often threaten rather than enhance these
packaged meals are delivered to tables by serving
goods.
robots, and to the showers, where something that
looks like a cross between an octopus and a car wash
bathes them carefully. Again, you observe that some 2
Coeckelberg (Coeckelbergh 2012) outlines a similar scenario as a
residents control the wheelchairs using a joystick or possible vision of the future of aged care in a paper of which I only
voice commands, while others appear to be moved became aware after drafting this one.
3
around at the initiative of the chairs themselves. In For a recent survey of such claims, see Vincze et al. (2015)).

123
AI & Soc (2016) 31:445–454 447

Third—and perhaps most controversially—I want to do justice to this body of thought here, a quick account of
argue that the introduction of robots into the aged care the main dialectic in the literature will help us to see that
setting is likely to transform aged care in accordance with a human welfare consists in much more than happiness.5
trajectory of development that leads towards this dystopian Of course, happiness is clearly a good thing and an
future even when this is not the intention of the engineers important component of well-being. However, it is equally
working to develop robots for aged care. clear that happiness is not the proper measure of the quality
Finally, I want to suggest that even when technology use of someone’s life. It would be an uncontroversially bad
is autonomous, as it is in at least some cases in the scenario way of caring for people, for example, to strap them to their
I have described, it may nevertheless remain problematic beds while they were asleep and then dope them up with
because of the ways in which technology embodies and mood elevating drugs or maintain them on morphine drips
establishes power relations between different groups of so that they were in a state of continuous ecstasy.
citizens and thus threatens respect for older citizens. For this reason, hedonistic accounts of well-being,
which place happiness or pleasure at their centre, are
unsatisfying. At the very least, what seems to matter is not
3 Happiness, well-being, and dystopia whether or not we are happy but whether or not we are
getting what we want. Are our lives going the way we want
The scenario I have just described is one in which the them to? Note that this is a different matter to whether or
residents appear to be happy while being cared for by not we think our lives are going the way we want them to
robots. This is perhaps the central feature of the scenario (Nozick 1974: 44–45). It is, possible, for instance, that we
that makes it science fiction. People at all stages of human think our life has a certain structure or valuable elements
life require human contact, both social interaction and when, in fact, it does not.
physical touch, for their psychological—and physical— However, as an account of what makes a human life go
well-being, and so it is exceedingly unlikely that people well, the satisfaction of desires or preferences is also
would flourish if cared for solely by robots. Nevertheless, extremely problematic. Some desires seem trivial, such that
it’s possible—although still, I think, unlikely—that some their satisfaction appears to contribute little to our well-
individuals, for instance, committed misanthropes or those being, while the satisfaction of other desires seems
with dementia severe enough that they are unable to dis- straightforwardly bad for us. If a person doesn’t want love,
tinguish robots from human carers, would be happy being family, beauty, or wealth but just wants to collect bottle
cared for entirely by robots. Thus, in order to address the tops, do we want to say that they have lived a successful
strongest possible case for the benefits of aged care human life if they die with a large bottle-top collection?6
robotics, I have outlined a scenario in which people are What if someone who is deeply depressed desires the
indeed happy in the care of robots. Indeed, I want to collapse of all those projects they had previously held to be
concede the possibility that the residents of this facility are, valuable? It is implausible to hold that the satisfaction of
in a non-trivial—if controversial—sense, happier than they any desire contributes to a person’s well-being—it also
would be if they were cared for by human beings in an matters what the desires are desires for.
alternative contemporary facility, where staff shortages and These problems are especially pressing for accounts of
low wages mean that human staff are often stressed and welfare that focus on the satisfaction of preferences
sometimes curt or rude. because of the phenomenon of ‘‘adaptive preferences’’
However, once I have acknowledged that the residents (Elster 1985, 109–110). Human beings are very good at
in this scenario are happy, my claim that it is dystopian adapting to even quite miserable situation and will typi-
may now seem puzzling. How can we say that people’s cally lower their ambitions to suit their circumstances. For
circumstances are bad when they are happy? this reason, we need to be extremely careful about con-
I hope that some readers will already share my intuition cluding that a person’s life is going well just because they
that this is not a future we should celebrate and strive for— are realising their desires.
even if it would be a happy one. However, in order to fully These two problems have therefore moved many
understand why this scenario remains a dystopia, we must philosophers to embrace what is called an ‘‘objective list’’
take a brief intellectual detour into the philosophy of theory of well-being (Arneson 1999; Griffin 1986; Rice
welfare. The question of how we tell when somebody’s life
is going well or whether they are harmed or benefited by 4
For a useful (if dated) survey, see Griffin (1986).
certain changes in their circumstances is absolutely 5
The account below roughly follows Parfit (1984: 493; subsequent
essential to social policy, as well as to the intellectual discussion).
foundations of economics, and so, it has attracted a great 6
A variation of a counter-example first suggested by Rawls (1971:
deal of philosophical scrutiny.4 While I will not be able to 432).

123
448 AI & Soc (2016) 31:445–454

2013).7 When we want to evaluate someone’s welfare, we respect as their content.9 For instance, polite and courteous
should consider the extent to which they have realised—or interactions with officialdom are part of recognition, while
perhaps simply have available to them—certain goods that granting citizens a vote in decisions that affect them is a
are objectively valuable. Are they healthy? Is their life free function of respect. Similarly, insults are an affront to recog-
from pain? Do they have friends and satisfying personal nition, while assaults involve the failure to respect their targets.
relationships? Have they adequate material comforts? Do Another way of characterising and distinguishing these goods
they have access to beauty? Do they enjoy the other goods is to identify their appearance in historical accounts of the
that make a human life meaningful and successful? Of nature of the ‘‘good life’’. For instance, recognition played a
course, the content of any such list is controversial, which central role in the Aristotelian virtue of ‘‘honour’’, which was
in turn has led some thinkers (Sen 1999; Nussbaum 2000, concerned with how one appears in the eyes of others, while for
2011) to conclude that we should privilege the capacity to Hegel (1977) it was foundational to subjectivity. In contrast,
obtain these goods over their possession, but this contro- Kant’s focus on the ethical requirement to relate to other
versy doesn’t seem especially irresolvable; if you ask human beings as members of the ‘‘Kingdom of ends’’
people what sorts of things contribute to a human life going emphasised the importance of respect.
well, there will usually be a remarkable degree of overlap Recognition and respect are important components of
in the lists that they come up with, if not in the precise human welfare because, as Aristotle (2004) (as well as many
rankings of goods on such lists (Rice 2013: 210–211). others) emphasised, human beings are fundamentally social
In any case, there are two goods that, I believe, are each animals. No human being can survive into adolescence—or
essential to any plausible list of objective goods, which flourish in adulthood—without a community. The nature of our
explain why the scenario I have described is dystopian. psychology is such that lack of human contact perverts us, even
First, there is an objective good, which I shall call ‘‘recog- where it is deliberately sought out. Social relations enter into
nition’’, which consists in the enjoyment of social relations that our very thoughts because the language we use is developed and
acknowledge us in our particularity and as valued members of nourished by a community. Our relation to that community and
a community. Second, there is an objective good, which I shall to its members is therefore central to our well-being. Depriva-
call ‘‘respect’’, which consists in social and political relation- tion of recognition, in particular, may have dramatic impacts on
ships wherein our ends are granted equal weight to those of a person’s subjective well-being and on their psychological and
others in the community. These goods are closely related and physical health. Lack of respect may be similarly corrosive but
are often enjoyed or absent together. However, they are in fact also involves the denial of a person’s moral worth regardless of
distinct.8 At a rough first approximation, we might think of whether or not they become aware of it.
recognition as a matter of the form of social relations and For current purposes, what matters is that these are both
goods that are constituted by certain types of relationships
7
An influential alternative involves introducing a requirement for between human beings. Machines lack both the interiority and
some degree of idealisation in the specification of the relevant desires. the capacity to enter into the rich sets of affective relations
Thus, for instance, we might say that people are well off when the
(which are constituted by mutual vulnerability and the par-
desires that they reflectively endorse when fully informed are
satisfied. Such accounts suffer from a tendency to collapse into ticular contingent features of human embodiment) necessary
versions of the ‘‘objective list’’ theory when placed under philosoph- to establish these ethical relations (Sparrow 2004). Thus,
ical pressure because it is difficult to quarantine accounts of the while clever design and programming might succeed in con-
reasonableness of desires from the worth of their objects.
8
vincing people that robots recognise their particularity and
Although the fact that relations between persons have this dual
respect their ends, they cannot in fact provide these objective
aspect is reasonably uncontroversial, both the precise way to make the
distinction and the most appropriate terminology by which to mark it goods (Sparrow 2002). People in the aged care facility I have
remain a matter of some controversy. The idea of ‘‘recognition’’ as a described are deprived of both recognition and respect by
distinct good was central to the philosophical debate about multicul- virtue of being looked after entirely by robots and for that
turalism, which took place in the 1990 s [see especially Taylor and
reason their welfare is jeopardised even if they are themselves
Gutmann (1992)] although the contrast with respect was not always
stated explicitly. Fraser (1995) comes close to making this distinction unconscious of this fact.10 Even someone with severe
as I make it here, although she cashes out the implications of a
concern for respect as a concern for the distribution of political and 9
economic opportunities. My account of recognition subsumes the first This can only be an approximation because recognition also admits
and third form of recognition distinguished by Honneth (1992) in his of the distinction between genuine and ersatz acknowledgement of the
justly influential account, while my concept of respect closely tracks worth of others.
10
the second form of ‘‘recognition’’ he identifies. In Nussbaum’s list of This is not to say that older persons are always treated with respect
capabilities, recognition is included within ‘‘affiliation’’, while respect and recognition by human ‘‘carers’’. However, where human beings
is most obviously represented as ‘‘control over one’s environment’’ don’t provide these goods, this is widely acknowledged to represent a
but is also represented in the concern with freedom and opportunity moral failing. As I discuss below, the claim that the use of robots in
that drives the focus on capabilities rather than a more determinate list aged care is inimical to respect is more controversial than the claim
of goods (Nussbaum 2011: 33–34). about recognition and I defend it further in the last part of this paper.

123
AI & Soc (2016) 31:445–454 449

dementia has a better quality of life when—as far as is pos- possible for people to stay out of any institutional setting—
sible—these relations are present, regardless of whether or not let alone one as ‘‘total’’ as the one that I have described—
they themselves are aware of them. Indeed, as I observed longer by developing robots that can support them in their
above, so central are these relationships to a good human life, daily lives and to remain in their homes.13 Second, they
that it is likely that only those deluded about their situation in will insist that rather than aiming to replace human beings
this home will in fact be happy.11 with robots in caring roles, their goal is to design and
Although I have not emphasised it here, there is a manufacture robots that will supplement and facilitate the
conceptual connection between respect and recognition and provision of good quality care by human beings: the future
the provision of the ‘‘care’’ that should be at the heart of of aged care will be ‘‘humans plus robots’’ rather than
aged care. As I have argued at length elsewhere (Sparrow ‘‘robots instead of humans’’. Third, they will agree that
and Sparrow 2006), robots cannot provide genuine care nobody should be forced to accept a robot carer when they
because they cannot experience the emotions that are don’t want one but argue that where people have con-
integral to the provision of such care. Another way of sciously chosen to employ a robot to assist in their care my
making the same point, though, would be to observe that points about the value of recognition and respect have little
genuine care affirms the worth and individuality of the weight (Borenstein and Pearson 2010: 286). In short, they
persons being cared for through the provision of recogni- will either deny that my scenario accurately anticipates the
tion and is guided by a concern for their wishes and pro- ends of their project or that it is necessarily dystopian.
jects founded in respect. For the remainder of the paper, I will address each of
these arguments in turn.

4 The best laid plans of engineers… 4.1 ‘‘Robots at home’’ or ‘‘robots in nursing
homes’’?
A world in which older people were cared for only by
robots might be a dystopia, then, even if the people being As I noted above, people have been talking about the
cared for were happy.12 Yet an argument that some pos- advent of robotic butlers ever since the dawn of robotics.
sible future is dystopian is neither here nor there if that Yet there are a number of reasons why this long-anticipated
future is highly unlikely to arrive. Given that I have already future has proved so elusive, which also suggest that robots
conceded that the scenario I describe above is science are much more likely to be successful in institutional
fiction, one might well wonder what its relevance is to the contexts rather than households at least in the first instance
real world of (the design of) aged care robotics? and probably for many years to come.
I don’t, in fact, believe that we are ever likely to reach a Despite decades of research—and despite much pro-
point where people are cared for entirely by robots, gress being made in the field recently—perception and
let alone where they are happy being so, not least because object recognition remain extremely difficult for robots. As
I’m cynical about the utility of robots in aged care for the a consequence, robots work best in structured environ-
foreseeable future (Sparrow and Sparrow 2006). However, ments, where their programmers can anticipate the range of
it is possible that I am wrong in this—indeed one presumes situations they will encounter. Robots handle the chaotic
that those advocating pouring funding into research into and the unexpected badly. Often robots ‘‘work’’ when the
aged care robotics believe that there is a good chance that I environment—including the options available to humans in
am wrong. Regardless, I want to suggest that, by clarifying interacting with the robot—can be structured to suit the
the logic of the development of these technologies, this capacities of the machine. Individual households tend to be
scenario reveals something important about the project of cluttered and chaotic environments, which are hard for
developing robots for aged care settings even if they are robots to function in, but also expensive and difficult to
never likely to fully realise their potential. modify in order to suit machines. Designing or modifying
Those committed to this project are likely, I suspect, to environments to suit robots is much more plausible for
object to this suggestion on at least three grounds. First, institutions, which are often already shaped with reference
they will insist that the goal of their research is to make it to the specific requirements of the institution. Institutions
also typically generate clearly defined tasks—such as
11
As I have argued elsewhere (Sparrow 2002), the ethics of delivering meals or medications to residents according to a
designing artefacts that encourage this delusion is problematic. schedule—that robots are more likely to be capable of
12
Vallor (2011) argues, with some plausibility, that it would also be
a dystopia in so far as this is a world in which (potential) caregivers
13
are denied the opportunity to cultivate important virtues and to benefit Again, an objective that is highlighted in both the EU-funded
from contact with the elderly. For some reservations about the general ACCOMPANY Project (see: http://accompanyproject.eu/) and
form of this argument, however, see Sparrow (2015). HOBBIT Project (see http://hobbit.acin.tuwien.ac.at/index.html).

123
450 AI & Soc (2016) 31:445–454

succeeding at, where they would fail in more challenging session while presenters struggle to get their PowerPoint
environments. slides to display attests, improvements to the theoretical
Moreover, aged care robots, especially robots capable of capacity of a technology do not necessarily make for a
lifting items, are likely to be expensive artefacts. Institu- better user experience. The history of technology is replete
tions are more likely to be able to afford to buy such with cases where technically superior technologies failed to
devices than individual households, not least because displace existing alternatives because they were more
institutions will benefit from economies of scale when expensive for too long and/or because the economic or
purchasing them. Any robot capable of playing a useful attitudinal costs involved in transitioning to the new tech-
role in aged care will also require sophisticated software, nology were too high. Credible predictions about the future
which is extremely likely to be updated regularly. If our of robots going ‘‘beyond the laboratory’’ must therefore
most sophisticated personal assistants to date—our mobile draw upon expertise in sociology, economics, marketing
phones—are any guide, the robots of the future will often and politics, as well as engineering.
be quirky and confusing and will function at far less than Perhaps more fundamentally, a scenario in which
their optimal capacity. Consequently, aged care robots are elderly persons remain in their own homes being cared for
likely to require a significant amount of technical support, by robots, with little contact with the broader community,
which it is again more plausible for institutions to purchase is not necessarily less dystopian than the picture of insti-
and/or access. Robots are also likely to suffer from rapid tutional care I painted above. As I have argued here, it is
obsolescence, as new versions are developed and ‘‘network easier to imagine a nursing home staffed entirely by robots
effects’’ render otherwise functional devices useless when than a robot in every home, but the real source of our
the infrastructure (batteries, cables, user manuals, qualified ethical unease about the former—the replacement of
service technicians, et cetera) required to maintain them human interaction with impersonal mechanical services—
disappears as a result of other users upgrading to new would also be present in the latter case.
models.
It is also worth pointing out that various ergonomic and 4.2 ‘‘Humans plus robots’’ or ‘‘robots instead
interior design changes, social interventions, and commu- of humans’’?
nications technologies will usually be both more effective
and much cheaper than any robot (Sorell and Draper 2014; In sketching a future involving an entirely robotic nursing
Sparrow and Sparrow 2006). Falls monitoring can be done home, I have suggested that the fundamental dynamic
with alarm pendants or motion detection systems. Telep- established by the development of robots leads to the
resence can be achieved with a laptop or mobile phone. replacement of humans in caring roles. However, many
Mobility may be facilitated through the addition of hand- roboticists see themselves as developing robots that will
rails, the replacement of stairs with ramps, and wheel- assist human beings in providing care and therefore to
chairs. The need for assistance with reaching for and lifting provide better quality care.
household items may be minimised by good ergonomics It is worth observing at this point that if a robot is
and universal design.14 meeting a real need in an aged care context then it is
For these reasons, I think it’s highly unlikely that meeting a need that could also be—whether or not it
we’ll see a robot to facilitate independent living in every actually is being—met by a human being. The possibility
home for at least the next several decades, if at all. If that robots might replace human beings in this context
robots for aged care do arrive, they will arrive first and therefore necessarily emerges from the project of devel-
flourish most in institutional contexts like the one I oping robots for aged care.15
described above. Thus, the key question is whether such robots will in
Admittedly, this is—at least in part—a criticism of the fact be used to substitute for human carers or to supplement
sort that I conceded at the outset might be rendered moot the care human beings can provide without reducing the
by progress in robotics. Yet the argument here is not number of human beings involved in caring for older cit-
entirely about the technical capacities of robots; it is at izens (Parks 2010). Again, this is not a question about the
least as much, if not more, about the utility and economics capacities of robots but about the economics of their future
of high-tech products as they enter the marketplace. As the use, if any.
inevitable delay at the beginning of many a conference
15
This is, perhaps, not so immediately obvious in the case of
14
It must be acknowledged that some of the changes required may be telepresence robots, which might be thought of as offering a new
very expensive in some homes and also that some of these changes, medium through which contact between people may occur. However,
where they are inexpensive, may also make it easier for robots to even in this case such robots clearly function to substitute for the
function in homes. physical presence of the other person.

123
AI & Soc (2016) 31:445–454 451

We can reliably anticipate that robots will come into carers.17 If robots won’t serve to replace human beings in
regular use in the aged care setting when they can per- aged care roles, then they won’t ‘‘solve’’ the problem posed
form, at a lower cost, tasks that would otherwise require a by the growing number of older citizens and the relatively
human being. The question then becomes whether the shrinking number of workers available and qualified to care
cost savings made possible by the substitution will be for them.
reinvested to achieve a higher standard of care and, in The argument that the introduction of aged care robots
particular, to ensure that older persons have the same (or won’t reduce the amount of human contact available to
a greater) number of opportunities for meaningful human older persons thus strikes me as at best naive, in the context
contact. of the economics of the age care setting, and at worst
Unfortunately, I believe it is naive to think that this will duplicitous.
occur. It is naive because of the economic pressures in the
age care sector, which motivate providers to cut costs
wherever they can. These are especially acute in countries 5 Respect and the autonomous use of robots
such as Australia, wherein age care is increasingly the
responsibility of private for-profit providers. These organ- Often, of course, one reason why people are happy is that
isations have a clear track record in Australia of cutting their lives are going well. Moreover, even on an objective
services to the bare minimum required to meet their legal list theory of welfare, there is room for people flourishing
obligations—and one suspects that the situation is little while choosing to pursue different bundles of those things
different elsewhere.16 Even where aged care remains part that are objectively good. For that matter, leading a life that
of a government-provided social safety net, there are at one has chosen for oneself is itself highly likely to feature
least three dynamics pushing towards reducing the cost of as an important good on any plausible list of such goods. If,
the provision of this service. First, government and chari- at some time in the future, people are happy being looked
table providers must typically ‘‘compete’’ with the private after by robots, then this might reflect the fact that they
sector or see funding flow to the latter and must make genuinely prefer this set of arrangements to being looked
similar cost cuts, where possible, accordingly. Second, the after by human beings. Indeed, there are at least two cir-
social welfare budgets from which age care is funded have cumstances in which it seems plausible that people might
themselves come under regular and repeated attack in well prefer robot carers over human carers. First, people
social democracies around the world over the last three may prefer care provided by robots where the needs being
decades as a result of larger-scale political and economic met involve intimate physical interaction, as in the case of
transformations. Third, the much discussed increase in the bathing or dressing, where individuals are often embar-
percentage of the population who are older citizens is rassed by having these needs addressed by human beings.
undercutting the tax base for the social provision of aged Second, as discussed above, they may prefer robots if they
care at the same time as it is increasing the need for such make it possible for them to stay in their own homes longer
care (United Nations, Department of Economic and Social rather than to move into an institution.
Affairs, Population Division 2013). These are all reasons to It might therefore be argued that where the choice to
think that, if they become available, aged care robots are embrace robot care is an autonomous one, even though the
likely to be used to substitute for rather than supplement robots themselves cannot provide respect, care by robots
care by human beings even in facilities in the not-for-profit may—in so far as it reflects the ends of the person being
and government sectors. cared for—be neutral with regard to respect (instead of
Indeed, the role played by claims about a ‘‘looming jeopardising respect), while the loss of recognition
demographic crisis’’ in the argument as to why we should involved may be a price that people are willing to pay to
be investing in robots for aged care (see, for instance, secure other goods that are more important to them.
Schaeffer and May 1999) suggests that ultimately at least There is some force in this objection; the idea that robots
part of the purpose of developing aged care robots is pre- might facilitate and enhance older persons’ autonomy is
cisely to make it possible for them to substitute for human
17
This goal is even made explicit in the announcement of a new
position paper on workforce issues by Aged and Community Services
Australia (ACSA), which quotes one of the authors, Adjunct
16
Striking accounts of the impacts of this dynamic in the Australian Professor John Kelly, as saying ‘‘We have to use robotics and
context are provided here: http://www.theglobalmail.org/feature/how- technology in clinical care to decrease—not in a huge way but
our-twilight-years-are-ripe-for-the-picking/73/ (accessed 15.1.15); possibly by 5 per cent—the amount of people we need by actually
and, here http://www.agedcarecrisis.com/yoursay/4611-no-staff-for- getting technology to do things’’. See Belardi (2015). New workforce
10-5-hours-per-day (accessed 15.1.15). My thanks to Linda Sparrow strategy for aged care. http://www.australianageingagenda.com.au/
for these examples. 2015/01/13/new-workforce-strategy-aged-care/ (accessed 15.1.15).

123
452 AI & Soc (2016) 31:445–454

indeed the most plausible argument in favour of the project sectors reinvigorated through the production of robots for
of developing robots for aged care (Borenstein and Pearson aged care will do exactly that.
2010; Sorell and Draper 2014).18 Finally, as this last observation implies, the implications
However, there are three grounds for thinking that this of aged care robotics for respect are more complicated than
argument is much weaker than initially appears. the claim that autonomous choice enhances respect
First, even if robots are compatible with respect, their suggests.
use occurs at a cost to recognition, which robots cannot The decision to use a new technology—and, in partic-
provide.19 One consequence—indeed a virtue—of objec- ular, to bring it into the home and use it daily—is not an
tive list accounts of welfare is that they allow for the insignificant matter; it has all sorts of ramifications, many
possibility that people’s own assessment of their best of which may only become obvious in retrospect and some
interests doesn’t settle the matter. Sometimes individuals’ of which may remain obscure even then. Philosophy and
choices will leave them worse off, even after taking into sociology of technology tell us that tools are not neutral.
account the goods associated with acting autonomously By foregrounding some possibilities for action and con-
and with other people respecting their choices. In the cealing or reducing others, they shape the ends of users
scenario I described above, residents have no face-to-face (Heidegger 1993; Winner 1986). Nor are the effects of
contact with anyone outside of the facility. While others technology confined to those who use them; technology
may disagree, I am inclined to think that this is a dystopian also places people in new relations with each other. Some
outcome even if it has the virtue that people are getting of these new relations are obvious, as when one comes to
what they want. rely on the staff on the IT support desk answering their
Second, people choose from amongst the options real- phones in order to be able to use one’s computer. Others
istically available to them. If industrialised societies invest are more subtle, as when people who are not on Facebook
in robotics in an attempt to meet the needs of older citizens miss out on invitations from their friends or those who are
rather than higher wages for workers in the sector or more not on Twitter have a different sense of the ‘‘events of the
innovative social reforms, which might integrate older day’’ to those around them. By placing people in new
persons and their care into the community, then in the relationships, technology alters the power relations
future older citizens may face the choice of being cared for between people (Winner 1986). Indeed, one of these new
by robots or living (or, indeed, dying!) with key needs relationships is precisely the fact that the choices that the
remaining unmet.20 It would hardly be surprising if they designers make regarding the design of technologies are
should choose to be cared for by robots in this context. now shaping the users’ ends and their relationships with
Moreover, because, as mentioned above, people tend to other people. All of which is to say that technology
adapt their attitudes to their circumstances, we should establishes a political relation between designers and users.
expect them, for the most part, to be happy with this Thus, when considering the impact of robotics on the
choice. Yet it would clearly be a mistake to use this pos- extent to which the ends of those being cared for are
sibility as a justification for the neglect of the policies that acknowledged to have equal weight to those of other citi-
might have allowed older citizens a genuine alternative. zens, it is not sufficient to ask whether the decision to use a
There is, I believe, a real danger that policymakers dazzled particular device is autonomous: it is also important to
by the prospect of engineering faculties and manufacturing think about who is designing robots and how the interests
of end-users are represented in the device and what role
they have played in its design.
18
This fact is also, I think, reflected in the recent enthusiasm in the Now it is true that these questions are not unique to
literature for Nussbaum’s ‘‘capacities approach’’ to social justice,
robotics. It is my firm belief that contemporary societies,
which acknowledges the value of autonomy by focusing on the
distribution of the capacity to achieve various ‘‘functionings’’ rather which are increasingly shaped by the social impacts of
than on goods understood as particular items or outcomes, as a lens technologies rather than by social and political movements,
through which to view the ethics of robots in aged care. See, also: should be paying a lot more attention to the questions of
Coeckelbergh (2012), Parks (2010), Sharkey (2014).
19
who is determining technological trajectories and what
This way of putting the matter risks implying that robots can
sorts of values are embedded in the technologies we
provide respect, which is not the case; rather, where individuals
choose to allow robots to be used to assist them, the decision of third embrace. Nevertheless, there are two reasons why the
parties to respect this choice provides the good of respect. impact of the introduction of aged care robots may involve
20
Again, it is clear that existing standards of care are already a more significant exercise of power over end-users than
inadequate in many, indeed arguably most, cases today. However, this other technologies. In so far as they are designed to
sad fact does not detract from the larger argument here; that claims
accomplish or facilitate actions within the daily life of
about the significance of older citizens’ willingness to be happy with
care provided by robots need to be treated with a certain degree of older persons—and in order to make this design task
cynicism, while the alternative is so dire. manageable—the design of such devices already

123
AI & Soc (2016) 31:445–454 453

necessarily involves a larger number of choices about what provide respect for older persons by allowing the options
sorts of activities or goals are important than many other available to them to be shaped by the design choices of
technologies.21 There are also prima facie grounds to think others.
that the pre-existing inequality between designers and users Inevitably, the claim that we are headed for a dystopia is
of robots for aged care are likely to be reflected in and itself a form of pessimism, which I said I wanted to abjure.
enhanced by the processes whereby aged care robotics However, the pessimism evinced by my analysis here is
come into use. At one end of this relationship, we have pessimism about the social and economic dynamics shap-
engineers at elite universities, well-educated government ing the provision of aged care now and into the future in
officials, and wealthy corporations; at the other end, we industrialised societies, rather than pessimism about the
have socially isolated older persons; it would hardly be skills and ingenuity of engineers. Correspondingly, those
surprising if the ends of the latter were granted less than who wish to pursue a more utopian vision for the future of
equal weight in this process. aged care than that I have outlined here would, I suspect, be
While a world in which persons in advanced old age well advised to think more about the real world of aged
autonomously chose to be cared for by robots might not care today and less about robots.
undercut respect, then, it is more likely that deferring to
this ideal will bring about a world in which older citizens Acknowledgments The research for this paper was supported under
the Australian Research Council’s Centres of Excellence funding
have little choice but to accede to receiving an increasing scheme (Project CE140100012). The views expressed herein are
number of services, which once would have been provided those of the author and are not necessarily those of the Australian
by human beings, from robots and suffer a net loss of well- Research Council. I would like to thank Professor Gesa Lindemann
being as a result. and Professor Gregor Fitzi for the invitation to attend the ‘‘Going
beyond the Laboratory’’ conference. I’d also like to thank my mother,
Linda Sparrow, and Catherine Mills for comments and discussion
during the process of drafting this manuscript.
6 Conclusions

Thinking about technological futures is always a fraught

matter: attempting to evaluate them ethically is even more References
fraught. I have suggested here that it would be a mistake to
let an initial pessimism about the capacities of robots pre- Aristotle (2004) The politics (E. Barker, Trans.). Oxford University
vent us from thinking critically about a possible future in Press, Oxford
Arneson RJ (1999) Human flourishing versus desire satisfaction. Soc
which robots have left the laboratory and entered aged care Philos Policy 16(1):113–142
facilities. Hypothesising a world in which engineers have Borenstein J, Pearson Y (2010) Robot caregivers: harbingers of
‘‘succeeded’’ in designing robots for aged care is also a expanded freedom for all? Ethics Inf Technol 12(3):277–288
useful exercise for diagnosing the ethical character of this Coeckelbergh M (2012) ‘‘How I learned to love the robot’’:
Capabilities, information technologies, and elderly care. In:
project. I have argued that paying attention to the objective Oosterlaken I, van den Hoven J (eds) The capability approach,
component of well-being reveals the importance of inter- technology and design. Springer, Dordrecht, pp 77–86
personal relations of recognition and respect, which robots Elster J (1985) Sour grapes: studies in the subversion of rationality.
cannot provide, to human welfare and, thus, the ways in Cambridge University Press, Cambridge
Fraser N (1995) From redistribution to recognition? Dilemmas of
which the applications of robots in the aged care setting justice in a ‘post-socialist’ age. New Left Rev 212:68–93
threatens to undercut these goods. Robots are much more Griffin J (1986) Well-being. Clarendon Press, Oxford
likely to be used in institutional settings than in individual Hegel GWF (1977) Phenomenology of spirit. Oxford University
homes, and a realistic understanding of the economics of Press, Oxford
Heidegger M (1993) The question concerning technology. In: Basic
aged care suggests that robots capable of performing rele- writings (Rev. and expanded ed). Harper, San Francisco
vant services will replace human workers in this setting and Honneth A (1992) Integrity and disrespect: principles of a conception
reduce the opportunity for these important goods. While of morality based on the theory of recognition. Polit Theory
some older persons may well embrace the use of robots in 20(2):187–201
Nozick R (1974) Anarchy, state and Utopia. Basic Books, New York
particular roles, overemphasising this possibility risks ren- Nussbaum MC (2000) Women and human development: the capa-
dering it a self-fulfilling prophecy, depriving those being bilities approach. Cambridge University Press, Cambridge
cared for of valuable social recognition, and failing to Nussbaum MC (2011) Creating capabilities: the human development
approach. Harvard University Press, Cambridge
Parfit D (1984) Reasons and persons. Clarendon Press, Oxford
Parks JA (2010) Lifting the burden of women’s care work: should
21
For an extended discussion of the various ways in which values robots replace the ‘‘human touch’’? Hypatia 25(1):100–120
are—and might be—embedded into the design of caring robots, see Rawls J (1971) A theory of justice. Harvard University Press,
van Wynsberghe (2013). Cambridge

123
454 AI & Soc (2016) 31:445–454

Rice CM (2013) Defending the objective list theory of well-being. Taylor C, Gutmann A (1992) Multiculturalism and ‘‘the politics of
Ratio 1(2):196–211 recognition’’. Princeton University Press, Princeton
Schaeffer C, May T (1999) Care-o-bot-a system for assisting elderly United Nations, Department of Economic and Social Affairs,
or disabled persons in home environments. In: Buhler C, Knops Population Division (2013) World population ageing 2013. ST/
H (eds) Assistive technology on the threshold of the new ESA/SER.A/348
millenium. IOS Press, Amsterdam Vallor S (2011) Carebots and caregivers: sustaining the ethical ideal
Sen A (1999) Development as freedom. Knopf, New York of care in the twenty-first century. Philos Technol 24:251–268
Sharkey A (2014) Robots and human dignity: a consideration of the van Wynsberghe A (2013) Designing robots for care: care centered
effects of robot care on the dignity of older people. Ethics Inf value-sensitive design. Sci Eng Ethics 19(2):407–433
Technol 16(1):63–75 Vincze M, Weiss A, Lammer L, Huber A, Gatterer G (2015) On the
Sorell T, Draper H (2014) Robot carers, ethics, and older people. discrepancy between present service robots and older persons’
Ethics Inf Technol 16:183–195 needs. In: 23rd IEEE international symposium on robot and
Sparrow R (2002) The march of the robot dogs. Ethics Inf Technol human interactive communication (IEEE RO-MAN 2014),
4(4):305–318 August 25–29, 2014, Edinburgh. http://hobbit.acin.tuwien.ac.at/
Sparrow R (2004) The turing triage test. Ethics Inf Technol publications/hobbit_roman.pdf. Accessed 21 Jan 15
6(4):203–213 Winner L (1986) The whale and the reactor: a search for limits in an
Sparrow R (2015) Imposing genetic diversity. Am J Bioeth age of high technology. University of Chicago Press, Chicago
15(6):2–10. doi:10.1080/15265161.2015.1028658
Sparrow R, Sparrow L (2006) In the hands of machines? The future of
aged care. Mind Mach 16:141–161

123