Children-Robot Friendship, Moral Agency, and Aristotelian Virtue Development

Abstract

Social robots are increasingly developed for the companionship of children. In this article we explore the moral implications of children-robot friendships using the Aristotelian framework of virtue ethics. We adopt a moderate position and argue that, although robots cannot be virtue friends, they can nonetheless enable children to exercise ethical and intellectual virtues. The Aristotelian requirements for true friendship apply only partly to children: unlike adults, children relate to friendship as an educational play of exploration, which is constitutive of the way they acquire and develop virtues. We highlight that there is a relevant difference between the way we evaluate adult-robot friendship compared to children-robot friendship, which is rooted in the difference in moral agency and moral responsibility that generate the asymmetries in the moral status ascribed to adults versus children. We look into the role played by imaginary companions (IC) and personified objects (PO) in children’s moral development and claim that robots, understood as Personified Robotic Objects (PROs), play a similar role with such fictional entities, enabling children to exercise affection, moral imagination and reasoning, thus contributing to their development as virtuous adults. Nonetheless, we argue that adequate use of robots for children’s moral development is conditioned by several requirements related to design, technology and moral responsibility.

Full text

Children-Robot Friendship, Moral
Agency, and Aristotelian Virtue
Development
Mihaela Constantinescu
1
*
†
, Radu Uszkai
2
†
, Constantin Vică
1
†
and Cristina Voinea
2
†
1
CCEA, Faculty of Philosophy, University of Bucharest, Bucharest, Romania,
2
Department of Philosophy and Social Sciences,
Bucharest University of Economic Studies, Bucharest, Romania
Social robots are increasingly developed for the companionship of children. In this article
we explore the moral implications of children-robot friendships using the Aristotelian
framework of virtue ethics. We adopt a moderate position and argue that, although
robots cannot be virtue friends, they can nonetheless enable children to exercise ethical
and intellectual virtues. The Aristotelian requirements for true friendship apply only partly to
children: unlike adults, children relate to friendship as an educational play of exploration,
which is constitutive of the way they acquire and develop virtues. We highlight that there is
a relevant difference between the way we evaluate adult-robot friendship compared to
children-robot friendship, which is rooted in the difference in moral agency and moral
responsibility that generate the asymmetries in the moral status ascribed to adults versus
children. We look into the role played by imaginary companions (IC) and personified
objects (PO) in children’s moral development and claim that robots, understood as
Personified Robotic Objects (PROs), play a similar role with such fictional entities,
enabling children to exercise affection, moral imagination and reasoning, thus
contributing to their development as virtuous adults. Nonetheless, we argue that
adequate use of robots for children’s moral development is conditioned by several
requirements related to design, technology and moral responsibility.
Keywords: imaginary companions, virtue ethics, Aristotle, social robots, children-robot friendship, human-robot
friendship, virtue, moral agency
INTRODUCTION
Advances in modern robotics have increased artificial agents’sensing and visual processing
capacities as well as mobility, pushing robots from industrial warehouses and factories to
unlikely new places, such as hospitals, therapists’offices, classrooms, and even homes (Darling,
2021). One of the most exciting avenues in robotics is the development of social robots relying on
deep-learning artificial intelligence that can take on tasks that were traditionally assigned to human
beings. For example, in recent years, there has been an increased interest in the design of social robots
for the companionship and care of children (Dawe et al., 2019), with examples such as Embodied’s
“Moxie,”SoftBank Robotics’“Pepper,”and Blue Frog Robotics’“Buddy.”Robots are intended to
assist and even replace teachers, nannies, friends and children’s therapists (Sharkey and Sharkey,
2010;Zhang et al., 2019;Di Dio et al., 2020;Pashevich, 2021).
In this article we explore the moral implications of children-robot friendships using an
Aristotelian framework. We argue that, although robot companions cannot be virtue friends,
Edited by:
Pekka Antero Mäkelä,
University of Helsinki, Finland
Reviewed by:
Michael Winter,
University of St. Thomas,
United States
Kamil Mamak,
Jagiellonian University, Poland
*Correspondence:
Mihaela Constantinescu
mihaela.constantinescu@
filosofie.unibuc.ro
†
These authors have contributed
equally to this work
Specialty section:
This article was submitted to
Ethics in Robotics and Artificial
Intelligence,
a section of the journal
Frontiers in Robotics and AI
Received: 19 November 2021
Accepted: 20 June 2022
Published: 03 August 2022
Citation:
Constantinescu M, Uszkai R, VicăC
and Voinea C (2022) Children-Robot
Friendship, Moral Agency, and
Aristotelian Virtue Development.
Front. Robot. AI 9:818489.
doi: 10.3389/frobt.2022.818489
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184891
ORIGINAL RESEARCH
published: 03 August 2022
doi: 10.3389/frobt.2022.818489

they can nonetheless enable children to exercise ethical and
intellectual virtues, and thus contribute to children’s moral
development. The reason is that the Aristotelian requirements
for true friendship apply only partly to children: unlike adults,
children relate to friendship as an educational play of exploration,
which is constitutive of the way they acquire and develop virtues.
We highlight that there is a relevant difference between the way
we evaluate adult-robot friendship compared to children-robot
friendship, which is rooted in the difference in moral agency and
moral responsibility that generate the asymmetries in the moral
status ascribed to adults versus children.
To develop this point, we look into the role played by
imaginary companions (IC) and personified objects (PO) in
children’s moral development and claim that robots play
similar parts in children’s lives as these fictional entities. In
particular, Personified Robotic Objects (PROs) can be
important assets in enabling children to exercise affection,
moral imagination, and reasoning, contributing to their
development as virtuous adults. However, this contribution to
children’s moral development is conditioned by meeting several
requirements for adequate use, such as 1) responsible design that
avoids positioning the robot companion as the equivalent of a
human companion, for instance by avoiding situations when the
robot reciprocates affection as if it were a true friend; 2)
technically embedding scientific features that add new
interesting challenges to kids’reasoning abilities and make use
of DIY toolkits with participative co-design; 3) proper ascriptions
of moral responsibility to those who may bear blame for improper
interaction between children and robot companions, such as
parents or tutors, developers, marketers, etc.
CHILDREN AND ROBOTS: PARTNERS IN
VIRTUE OR IN VICE?
In this section we look into possible moral benefits and harms
that interaction with robots might bring to children, in particular
when this type of interaction is viewed as a form of
companionship or friendship. We start with results from
practice, which suggest that robots are accepted by children as
they develop relationships based on trust and a common ground.
Despite the potential benefits in children-robot interactions that
empirical studies point towards, ethical problems still abound.
We continue with theoretical accounts raising concerns, among
others, over the potentially deceiving nature of robots engaging in
relationships with humans.
Children-Robot Interaction in Practice
Social robots are more than toys. Unlike usual toys, robots can
move, interact, and communicate with children in complex ways.
They can adjust their behaviour and responses to accommodate
children’s specific needs and interests. In a world of over-crowded
classrooms and overly busy parents, social robots that can look
after children and provide companionship are indeed something
to look forward to. This is why there is a growing interest in how
children interact with robots in practice, as shown by the flurry of
empirical research on this topic.
In an important study, Kahn et al. (2012) found that children
attribute feelings, intelligence and interests (mental states) to the
robot used in the study (Robovie) and they even see it as a
potential partner or even friend. Indeed, the more similar the
robot is to a human being, the easier it is for children to attribute
mental states to the artificial agent, which means that robots can
easily be considered part of the friends’circle (Beran and
Ramirez-Serrano, 2010;Manzi et al., 2020). Explaining to
children that robots are “mere mechanisms”and not alive,
does not seem to have a profound impact on how children
attribute mental states to robots (Bartlett et al., 2004). But this
does not mean that children are easily fooled by robots into
considering them moral agents. Interestingly enough, children
are not very eager to grant robots liberties or rights (Breazeal
et al., 2016), despite attributing mental states to them. All of these
studies seem to suggest that children see robots as semi- “moral
others”(Nomura et al., 2016) which presupposes, some claimed,
the emergence of a new ontological category (Kahn et al., 2004;
Kahn et al., 2011;Kahn et al., 2012), as robots are seen by children
as “neither alive nor not alive, but something altogether
different.”(Severson and Carlson, 2010).
If indeed children perceive robots as potential partners, then it
is no wonder that robots enhance learning processes (Kanda et al.,
2004;Verner et al., 2016). Autonomous teacher-robots are
considered credible by children, and they are successful in
producing behavioural learning (intent to enact the
recommendations made by the robot) (Edwards et al., 2016).
Children can learn from robots especially when they can attribute
to robots the ability to think and make decisions (Brink and
Wellman, 2020). For example, robots can stimulate children’s
curiosity and creativity (Ali et al., 2019a;Gordon, Breazeal, Engel,
2015). A scoping review confirms the potential benefits of social
robots for learning, but also stresses the challenges–both
technological and pedagogical–that must be addressed before
introducing artificial agents into every classroom (Belpaeme et al.,
2018). All in all, robot teachers are more effective than computer-
based tutoring, probably due to the embodied nature of the
experience (Belpaeme et al., 2013).
Another field where robots seem to be effective is healthcare.
Robot Assisted Therapy is a promising field in paediatrics, as
robots have successfully been used to provide companionship and
alleviate loneliness for children during therapy and hospital stays
(Blanson-Henkemans et al., 2012;Belpaeme et al., 2013). But
nowhere have results been more promising than in robot therapy
for autism spectrum disorders (ASD) (Coeckelbergh et al., 2016).
One study shows that children with ASD interacted more with a
social robot than with a human interlocutor—a fact that can be
partly attributed to children’s excitement and curiosity regarding
the robot (Kim et al., 2013). Yet other studies found that social
robots facilitate the learning of some social rules in children with
ASD (Ali et al., 2019b;Zhang et al., 2019). Another review points
towards the fact that robots do indeed elicit more engagement
and also novel social behaviours especially from children and
teenagers with ASD (Scassellati, 2007;Scassellati et al., 2018).
Researchers do not know why children with ASD are more
responsive towards robots, but they speculate that amongst the
reasons might be the fact that interaction with robots eliminates
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184892
Constantinescu et al. Children-Robot Friendship and Virtue Development

the stress of social interaction present in communication with
other human beings. In other words, interaction with robots
means less social overstimulation for children, as robots offer
more predictable and reliable responses than humans do
(Scassellati, 2007;Darling, 2021).
There is nonetheless a lack of longitudinal studies on how
robots influence children’s social and emotional development.
One reason is that today’s robots created especially for children’s
care are not made for interactions that are longer than a couple of
weeks or months; it is thus difficult and maybe too early for
researchers to track children-robot interaction over the years
(Pashevich, 2021). Another issue that complicates the study of
children-robot interaction is that advanced robotic technologies
are still very new, expensive, limited and not robust enough to be
used in day-to-day life for long periods of time, such as years.
Likewise, large-scale empirical studies involving such new and
emerging technologies might face a barrage of objections from
research ethics committees. Moreover, many studies mention the
“novelty effect”of interacting with robots, which basically boils
down to the fact that most children enrolled in the studies are at
their first encounter with complex robots and this might explain
their mostly positive reactions to artificial agents (Michaelis and
Mutlu, 2018). Furthermore, robotic technologies are very diverse,
both in their functions, as well as in their appearance. The context
(place and culture) in which the interaction takes place, as well as
the intervention the robots are used for (teaching, therapy,
nursing etc.) might all have a say in shaping children-robot
interactions. This is why even if empirical results are
promising, they are not enough, by themselves, to dispel the
theoretical concerns sketched below.
It is All Fun and Games Until Someone Ends
Up Vicious
More and more worries regarding the potential pitfalls of human-
robot interaction have been raised in the last few years. Among
them, taking robot companions to be friends has been discussed
as negatively affecting human users’moral, emotional, and social
abilities, because: 1) considering the robot to be a genuine friend
is deceptive (Elder, 2015;Nyholm 2020;Sharkey and Sharkey
2021), 2) undermines the value of social connections (Jecker,
2020;Prescott and Robillard, 2021;Sharkey and Sharkey, 2021),
3) does not contribute to the development of virtue, but only of
vice (Nomura et al., 2016;Sparrow, 2016), and 4) leads to moral
deskilling (Vallor 2016). In this section we focus on these
concerns and highlight especially the dangers that introducing
social robots into children’s lives might bring about for children’s
moral development.
The most pressing issues in adult-and children-robot
interactions have to do with the deceiving nature of the
relationship (Sparrow and Sparrow, 2006;Elder, 2015;
Danaher 2020;Sharkey and Sharkey, 2021). As long as robots
only simulate the reciprocation of friendship, treating them as
friends would only amount to a simulacrum of friendship
(Nyholm, 2020). In other words, social robots only feign
human socio-emotional reactions and behaviours, thus any
goodwill or affection towards them would be one-sided
(Matthias, 2015), leading to self-deception (Navon, 2021).
According to Sharkey and Sharkey, deception arises whenever
“the appearance and the way that a robot is programmed to
behave, creates, for example, the illusion that a robot is sentient,
emotional, and caring or that it understands you or loves you”
(2021, 311). There are still ongoing debates on what deception is
and under what circumstances it arises, for example, whether it
requires intention or not (Sparrow, 2002,2016;Elder 2015;
Matthias, 2015;Danaher, 2020), but under all views, deception
is morally problematic in view of its effects on the deceived or on
society at large (Sharkey and Sharkey, 2021).
In the case of children, who are an epistemically vulnerable
group, deception could further result in undermining the value of
social connections (Jecker, 2020;Prescott and Robillard, 2021;
Sharkey and Sharkey, 2021). Since growing up with artificial
companionship would be easier and less challenging than trying
to develop a friendship with a human peer, one might argue,
along the lines of Turkle (2017) that constant interactions with
robots might disincentivize children (and, as a result, future
adults) from pursuing relationships with other humans. In a
similar vein, (Sharkey and Sharkey 2010;Sharkey and Sharkey
2021), drawing upon research from developmental studies on
attachment, argue that robot companionship (especially in the
early stages of childhood) could end up harming children by not
equipping them properly for socialising with other humans.
Parents could also be “fooled”into believing that the robots
meet the emotional demands of children. This, in its turn, could
lead to robots being allocated social roles for which they are unfit.
Just like some schools might make a push for having robot
teachers in a time poised with staff shortages (Sharkey, 2016;
Sharkey and Sharkey, 2021) parents might actively “push”
children into treating their robot as a friend. Such a
relationship has the potential to damage children both
physically (a robot friend might not be able to adequately
distinguish all the time between a child’s adequate use of
scissors) or emotionally (Sharkey, 2016;Sharkey and Sharkey,
2021). The deception argument in the case of children emphasises
the fact that virtue cannot be developed and exercised outside the
scope of society, but only when interacting with other human
agents around them. Then, with less social connections with
human beings and more connections with social robots that
apparently satisfy children’s social and emotional needs, kids
might end up having less opportunities to develop their moral
character.
An additional worry regarding true friendship with robots
relates to the fact that human-robot interaction might not
contribute to the development of virtue, but only of vice in
the case of children. On the one hand, take, for example,
Sparrow’s intriguing cases of kicking robot dogs (2016).
Kicking a robot dog is morally problematic not because of the
impact that such an action might have on the object, but because
doing it reveals something about our dispositions: we have a cruel
and irascible moral character. In other words, the way in which
we interact with robots gives us a chance to show whether we have
a vicious moral character. But because robots do not normally
“object”or respond to abusive behaviour, children might be
incentivized to experiment cruel behaviour on them. This
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184893
Constantinescu et al. Children-Robot Friendship and Virtue Development

might desensitise or habituate children with vices, which is
especially worrisome if, as Nomura et al. (2016) showed,
children might derive pleasure from abusing their robot
companions.
On the other hand, if people act virtuously towards robots, this
will not prove that those people are virtuous agents. This is due to
the fact that robots are deprived of any moral status; thus, there is
no one there to take advantage of that kindness. The crux of this
asymmetry rests upon the role which practical wisdom plays in
vicious or virtuous behaviour. While vicious behaviour does not
need to be guided by practical wisdom, virtuous behaviour does.
It is for this reason that, according to Sparrow (2021, 5), “a person
who possessed practical wisdom would know that and would also
therefore realise that “kindness”towards robots does not realise
the goals—the improvement of the welfare of humans and
(perhaps) other sentient creatures—towards which kindness is
oriented. In the absence of the exercise of practical wisdom, virtue
is impossible and what would ordinarily count as the exercise of a
virtue—the demonstration of a kind disposition—fails to do so.”
Another possible detrimental effect of taking robot
companions to be genuine friends relates to children’s moral
deskilling. This worry has been taken up by Shannon Vallor,
(2015);Vallor, (2016), who claims that robot companions could
‘morally deskill’and thus prove to be “morally debilitating”for
people in their capacity to develop virtue friendship with other
human beings. According to Vallor, our moral skills (and this is a
point of utmost importance to children, as prospective moral
agents) are “typically acquired in specific practices which, under
the right conditions and with sufficient opportunity for
repetition, foster the cultivation of practical wisdom and moral
habituation that jointly constitute genuine virtue”(2015, 109). As
a result, in order to develop moral skill and cultivate phronesis, a
child needs to be exposed to 1) models of skilful practice, 2) the
proper institutional context (namely the cultural and social
norms) which reinforce such a skilful moral practice, 3)
possess proper basic moral motivation, 4) possess the
necessary cognitive and emotional resources, and 5) and
environment with a constant feedback loop that gives us the
opportunities to habituate and learn from failure (2015, 110).
While the possession of conditions 3) and 4) might simply rest
on the general moral and cognitive upbringing of the child, for the
remainder of them acquisition becomes more problematic in the
context of robot friendships. Take, for example, the role that the
exercise of patience (Turkle, 2017;Nyholm and Frank, 2017) has
in fostering genuine relationships and virtue friendship with
other human beings. While children will search for instant
gratification in relation to robot companions, the latter would
need to possess practical wisdom and a moral personality to
understand why patience is important in developing a
relationship with another human being and how that is
fostered by our social and cultural institutional context.
Most of the above theoretical concerns share an Aristotelian
understanding of moral development and virtuous life. In this
framework, real friendship is an important element in the
acquisition and development of virtue. Given that robots
cannot at the moment but simulate human mental and
emotional capacities, most researchers conclude that friendship
with robots can only be an illusion (Sparrow and Sparrow, 2006;
Elder, 2015;Turkle, 2017;Sharkey and Sharkey, 2021).
In what follows, drawing on the Aristotelian conception of the
value and role of friendship and the broader framework of virtue
ethics, we build the case that robot companions might
nonetheless have a positive impact on the moral development
of children, given the difference between the way we evaluate
adult-robot friendship compared to children-robot friendship.
ROBOT FRIENDSHIP, MORAL AGENCY,
AND VIRTUE ETHICS
Most of the current accounts concerned with friendship between
humans and robots are grounded in the Aristotelian account of
friendship and virtue (Elder, 2015;Elder, 2017;de Graaf, 2016;
Danaher, 2019;Nyholm, 2020) or at least take it into
consideration (Ryland, 2021). As a result, the focus of
contemporary debates is on delineating the criteria for what
Aristotle calls “virtue friendship”and see whether human-
robot friendship satisfies them. With but a few notable
exceptions (Danaher, 2019), most researchers argue that
current and foreseeable robots are unable to be our virtue
friends (Elder, 2015;de Graaf, 2016;Nyholm, 2020). However,
adopting different theoretical perspectives allows some to confer
human-robot friendship a relevant moral role (Gunkel, 2018;
Marti, 2010; Ryland, 2021) and reject the standard Aristotelian
account as being too demanding. Yet others highlight that such
alternative contemporary accounts endorse “some fairly watered-
down and less interesting sense of friendship”(Nyholm, 2020:
Sect. 5.7).
In this article we follow the mainstream approach and adopt
the standard Aristotelian view of friendship. The main reason for
taking this stance is that we consider it is still robust and relevant
today, in particular because the broader framework of virtue
ethics has already proved useful in inquiries over moral status,
moral agency and moral responsibility in the fields of machine
ethics (Wallach and Allen, 2008;Howard and Muntean, 2017)
and Human-Robot Interaction (Cappuccio, Peeters and
McDonald, 2020;Peeters and Haselager, 2021), and even more
when applied to robotic AI systems (Hakli and Mäkelä, 2019;
Coeckelbergh, 2020;Sison and Redín, 2021;Constantinescu et al.,
2022). Another reason is because we consider that alternative
contemporary accounts of friendship did not (yet) provide
sufficient grounds for us to give up the standard Aristotelian
account, but rather to annotate it.
We discuss below the relevance of moral agency and moral
status for human-robot friendship by drawing on Aristotelian
virtue ethics and contemporary criteria for friendship rooted in
the standard Aristotelian view. Our main claim is that
Aristotelian criteria are robust enough for inquiries over adult-
robot friendship but have some limitations when applied to
children-robot friendship. These limitations can be explained
through the differences in virtue, moral agency, and moral
responsibility of children, compared to adults. We find the
asymmetry in moral responsibility and moral agency between
children and adults, and between children and robots to be of
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184894
Constantinescu et al. Children-Robot Friendship and Virtue Development

particular interest when evaluating the possible relation of
friendship between children and robots.
Aristotelian Criteria for True Friendship and
Characteristics of Good Friends
As laid down in Aristotle’sNicomachean Ethics
1
(books VIII-IX),
friendship (Gr. <philia>) is of three species, corresponding to the
three reasons that make someone worthy of love (VIII, 3): 1) for
utility, in view of gaining some personal benefit, 2) for pleasure, as
the parties are guided by feelings in enjoying each other’s pleasant
presence, and 3) for virtue or for good (character), between “those
who are alike in their virtue: they each alike wish good things to
each other in so far as they are good, and they are good in
themselves”(VIII, 3, II56b5-10). While the first two species of
friendship are imperfect because they are instrumental, the third
one is considered by Aristotle to be true or complete friendship
because it is pursued for itself.
Contemporary research concerned with human-robot
friendship (Danaher 2019;Elder, 2015,Elder, 2017;de Graaf,
2016;Nyholm 2020;Ryland, 2021) have picked up this
Aristotelian view of friendship and proposed various sets of
criteria for perfect or virtue friendship (see Danaher 2019 and
Ryland, 2021 for classifications of main conditions referenced in
the literature). These conditions include, for instance, mutuality,
authenticity, equality, empathy, shared life, or associative duties.
While we consider the various conditions put forward to be
relevant and well-grounded in the Aristotelian account of
friendship, we think that they tend to ignore or place too little
emphasis on the characteristics of good friends as depicted by
Aristotle. This has a detrimental effect on the way we engage with
evaluations of human-robot friendship, leading to failure to
acknowledge the relevance of virtue, moral agency, and moral
responsibility for friendship within the broader account of
Aristotelian virtue ethics.
In Aristotelian virtue ethics, the right or virtuous action in a
particular context is the action that a person who acquired virtues
would choose to do: “A virtuous action in certain circumstances is
what is required in those circumstances and what a virtuous
person would do in those, or relevantly similar, circumstances”
(Crisp, 2015, p. 270). A right or virtuous action can thus only
result from a virtuous character (NE, 1106b18-24), as “there is no
such thing as an objectively virtuous action in itself, considered
independently of the person who performs it”(Sison and Ferrero,
2015: 86). It takes practice and habituation to acquire and develop
the necessary stable and enduring states or dispositions (hexeis)of
character that enable agents to correctly grasp the right thing to
do in specific circumstances, that is, ethical virtues (ethike areté).
Furthermore, ethical virtues are developed through the exercise of
reason and are made possible through the enabling role of
intellectual or dianoetic virtues (dianoetike areté), which
include knowledge of the necessary and eternal (the virtues of
episteme, nous, and sophia) and knowledge of the contingent and
contextual (the virtues of techné and phronesis)(Irwin 1999;
Mureșan, 2007;Meyer 2011;Crisp 2018). Thus, ethical virtues
rest on rational choice or deliberation (prohairesis) (NE 1106b36
and 1139a22), and are guided by practical wisdom (phronesis),
which enables agents to rightly deliberate across various contexts
of action (Irwin, 1999).
When we relate the broader claims of virtue ethics to
characteristics of good friends, we find it important to
highlight that to be a friend in the virtue friendship sense one
needs to be good, that is, virtuous (Nyholm, 2020). This happens
because virtue friendship takes place between good or virtuous
people, who lead a virtuous life. It is precisely because of their
being good or virtuous that such people are able to commit to
virtue friendship (NE VIII, 4, II57a30). They show friendship
towards each other “because of what they are, not for any other
incidental reason”(NE VIII, 3, II56b, 10–15). Interestingly,
Aristotle notes that friends involved in virtue friendship are
like alter-egos, as a good friend is another self of the good
person (NE IX, 89, II69b5): virtuous people become friends
and love one another for their characters (Crisp, 2018). As
Aristotle notes, friendship consists more in loving than being
loved, rendering loving the virtue of friends (VIII, 8, II59a35). For
the good person acts towards the good or the virtuous, having in
mind the sake of their friend, even if this means neglecting their
own interest (IX, 7, II68a35). As a result, empathy is an important
condition for virtue friendship (II66b34), understood as caring
for the good of the other, grieving and joying along with them. It
is through empathy that the friend becomes another self.
Furthermore, the process of developing or training virtues takes
place when good people live in each other’s company (IX, 9,
II70a10), because they become better and improve each other
through their activity (IX, 12, II72a10). True friendship requires
familiarity and time (VIII, 3, II56b, 25), experiencing interaction in
various situations. This is why friendship is both a virtue and
involves virtue (Crisp, 2018). This further explains why friendship
is also mutual and rests on rational choice: friendship is a state
rather than a feeling, a state that “makes people wish good things to
those they love, for their sake”(VIII, 5, II57b30). One important
implication of the Aristotelian framework is that true friendship is
between equals (VIII, 5, II58a). There can be, however, particular
species of genuine friendship involving superiority, such as that
between father and son or older to younger (VIII, 7, II58b10). This
means that the type of equality required between parties engaged in
virtue friendship involves equality in moral status (Nyholm, 2020).
At this point we seem to run into an obvious obstacle for
human-robot friendship: despite some openness to grant robots
moral standing (Gunkel 2018), most researchers would agree that
there is a difference in the moral status ascribed to robots as
opposed to human beings. Currently, the standard (and only)
entities that are ascribed full moral agency and moral
responsibility, which are the building blocks of what we call
moral status, are adult human beings (Hakli and Mäkelä, 2019).
Moreover, researchers highlight the incapacity of robotic AI
systems to be virtuous and thus take part in virtue friendships
(Nyholm, 2020;Constantinescu and Crisp, 2022). The discussion
over human-robot virtue friendship seems to ultimately revolve
around the moral status that we ascribe to robots and, relatedly, to
the possibility of them being moral agents. Similar to other
researchers adopting the standard Aristotelian view, we hold
that human-robot friendship cannot be considered virtue
friendship. However, in reaching this conclusion we do not
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184895
Constantinescu et al. Children-Robot Friendship and Virtue Development

weight human-robot friendship against a specific set of Aristotelian
criteria, but instead highlight that robots are unfittobevirtue
friends because they lack the required moral agency to make them
part of our moral community, that is, agents capable of virtuous or
vicious action for which they can receive blame or praise.
But what happens when neither robots, nor humans engaged
in human-robot interaction are full moral agents? Is it possible to
speak of virtue friendship in this case, given the faded difference
in moral status? We continue with a few remarks over the
implications of moral status and moral agency in the case of
children-robot robot friendship, as opposed to adult-robot
friendship.
Robot Friends, Moral Development, and
Moral Agency: Childhood Versus Adulthood
Building on Aristotelian virtue ethics, we would like to highlight
that there is a difference between the way we evaluate adult-robot
friendship compared to children-robot friendship. This is rooted
in the difference in moral agency and moral responsibility that
generate the asymmetries in moral status ascribed to adults versus
children.
Contemporary research discusses two main conditions to
ascribe moral agency and moral responsibility to an entity,
which are grounded in Aristotelian virtue ethics ( Coeckelbergh,
2020;Constantinescu et al., 2021;Sison and Redín, 2021): the
freedom and epistemic conditions. These conditions pick on
Aristotle’s discussion over criteria to ascribe blame or praise to
an agent depending on their acting viciously or virtuously,
introduced in his Nicomachean Ethics (Book III parts 1 to 5
and Book V parts 8 and 9). In the Aristotelian framework,
conditions for moral agency and moral responsibility are
inherently intertwined with conditions for virtue of character,
for it is only when individuals act as a result of their virtue or
vice that we might hold them praise- or blame-worthy (Meyer,
2011). Drawing on these Aristotelian distinctions, in order to be a
moral agent and bear moral responsibility, one needs to be able to
1) causally generate an outcome while 2) acting freely, uncoerced,
and 3) be knowledgeable of the contextual circumstances of their
action, following 4) deliberation based on rational choice, involving
reason and aforethought (for a broader discussion see
Constantinescu et al., 2022). Aristotle’sdiscussionoverthese
four criteria highlights that children are not yet moral agents
because they lack deliberation (prohairesis), which is a
constitutive condition of moral agency and moral responsibility,
requiring agents to be able to act based on rational choice.
This difference in moral status generates important differences in
the moral development of children engaging in friendly interactions
with robots, as opposed to adults engaging in friendship with robots.
On the one hand, when adults direct virtuous behaviour towards
robot companions, they seem to fail to understand that robots are
not appropriate targets for kindness, for instance (Sparrow, 2021). As
a result, adults engaging in apparently true or virtue friendships with
robot companions show that they lack the required practical wisdom
that guides morally responsible agents. On the other hand, when
children direct virtuous behaviour towards robot companions, we
might rather regard this as a learning process towards becoming
virtuous. Because children are not yet full moral agents, we do not
(and should not) hold the same expectations of them when it comes
to developing forms of friendship with robot companions. When
children bond with robot companions and direct virtuous behaviour
towards them, the reason is not because of failing to grasp the
context and failing to exert practical wisdom. This rather happens
because childhood is the age where individuals are trained to acquire
and to develop virtues. In other words, children do not yet have the
life experiences necessary for developing practical wisdom.
As a result, we consider that robot companions might have a
positive impact on the moral development of children. Children-
robot friendship is not and cannot be virtuous because both
parties lack the moral agency required for a virtuous moral status
normally ascribed to adult human beings, as laid down by the
Aristotelian theory of moral responsibility. However, Aristotle
himself notes that, as concerned with the young, friendship keeps
them from making mistakes (VIII, 1, II55a15), despite the fact
that the young only form friendships for utility and for pleasure,
because they “live in accordance with their feelings”(VIII, 3,
II56a30-35). Though not true or perfect friends, robot
companions envisioned, along Aristotelian lines, as friends for
pleasure or utility, might contribute to children’s ability to acquire
and develop virtues, for example through the characteristic
activity of childhood, that is, play. If robot friends contribute
to children becoming virtuous, this could further have a positive
impact on the way children exercise moral deliberation on their
path toward becoming morally responsible adults. Nonetheless,
this contribution to children’s moral development is conditioned
by meeting several requirements for adequate use. We discuss
these issues in the next section of our article.
PERSONALISED ROBOTC OBJECTS,
FRIENDSHIP AND CHILDREN’S MORAL
DEVELOPMENT
Friendship has received a great deal of attention, but its
paradigmatic example is adults’relationship. Rarely has the
issue of friendship between children and for children been in
the spotlight, because “the general attitude tends to be a “bus
theory”of friendship: do not worry if you miss one, another will be
along in a minute”(Healy 2011, 442). This attitude towards
children’s friendship has probably been informed by a
misreading of Aristotle’s view on friendship and its importance
for children. Indeed, children are not capable of virtue friendships
because they do not really possess virtues as relatively stable traits
of character (Kristjánsson, 2020). But, as we previously showed,
while Aristotle mentions that children can only form friendships
for utilityand pleasure, he nonetheless acknowledges the formative
dimension of these imperfect relationships for the development of
children’s characters: “the young need [friendship] to keep them
from error”(NE 1155a). Childhood is the time when individuals
learn to become virtuous and develop as future moral agents.
Friendship, even in imperfect kinds, such as for pleasure or utility,
is constitutive of children’s moral development.
In this section we draw an analogy between social robots and
imaginary companions (IC) and personalised objects (PO) to
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184896
Constantinescu et al. Children-Robot Friendship and Virtue Development

show that “imagined”and one-sided relationships of friendship
are not uncommon for children. Children have always used their
imagination to come up with new play friends, as is the case with
IC and PO (such as stuffed toys). What is more, IC and PO help
children learn about social relations, affection, care, and
responsibility - which is what robot companions can do well.
We position robot companions as Personalised Robotic Objects
(PROs) and explore the way they might enable children to
exercise primarily their intellectual virtues, and secondarily
their ethical virtues, thus contributing to children’s moral
development. We further develop several conditions that need
to be met so that adequate use of PROs may indeed contribute to
the exercise of virtues through childhood.
New Kids on the Block Practice Their
Virtues
Children have always had imaginary companions. Scientific
research on imaginary companions (IC) evolved during the
last decades from a very restrictive position–IC are a sign of
possible mental disorders–towards a more inclusive, less
medicalised one: ICs are common and normal, even beneficial
to the psychological and social development of children. A
longitudinal study from 2004, so before the advent of social
media and smartphones, found out that 65% of 7-year-old
kids had at least one imaginary friend at some point in their
childhood, and 28% of 12-year-old children still have them
(Taylor et al., 2004, 1182–83). ICs are for impersonation or
“pretend play”; this “fantasy behaviour”is not detrimental,
there is no evidence that having IC will produce negative
personality traits (Taylor et al., 2004, 1182). It seems that in
this aspect kids with IC are not different from those lacking a
penchant for fantasy. Newer approaches to this topic, even when
investigating the similarity between IC and hallucinations,
recognize the fact that “creation of fictional characters and the
generation of imaginary friends arguably share a feeling of
distributed agency paired with knowledge of the subjective
source of these creative acts”(Fernyhough et al., 2019, 8).
Furthermore, children with IC tend to outclass the less
imaginative ones in the domain of mental state reasoning and
social competence, even though it is debatable whether having IC
betters one off or IC is a result of one’s advanced abilities (Davis
2020, 380–81). Research shows that kids are very serious about
their play. This is why children tend to treat imaginary
companions as real-life, flesh-and-bone friends (Davis 2020, 382).
Furthermore, personified objects (PO–that is, tangible
imaginary companions, such as stuffed teddy bears), as fantasy
predispositions, are isomorphic with IC, as the activity involved is
the same: creating another entity for the sole purpose of
interacting with that presence, be it material or just mental
(Davis 2020, 376). The difference between the two rests in the
infinite possibilities of mental design of IC compared to the
limited set of shapes and behaviours of PO, such as social
robots or interactive toys. It is more challenging to design and
attribute a mind to a complex still lifeless machine that already
has a given form with its functions and affordances. Beyond this
difference, both IC and PO serve the purpose of exercising
children’s moral imagination which, in its turn, is constitutive
in the process of developing the intellectual virtue of practical
wisdom. Moral imagination is “the capacity to envision given
contexts from multiple, even incompatible, frames of reference to
ensure a broadened moral lens with which to approach and assess
lived experience”(Fletcher, 2016). The exercise of moral
imagination is especially important in childhood as it helps
children take note of and distinguish the ethically relevant
features in different contexts. What is more, moral
imagination is necessary for imagining situations not yet
encountered in experience and to compare them to alternative
situations. In other words, moral imagination is a sort of
“anticipatory moral appraisal”that helps people confronting
moral dilemmas to imagine alternatives that are practically
and ethically feasible and sound (Werhane, 1999).
Social robots could, in a sense, play a very similar role in
children’s lives as imaginary companions and personified objects
do. Children tend to imagine and project human-like
characteristics to all sorts of objects, be they inanimate or just
imaginary. The worry that we developed in the previous sections
is that children might do the same with social robots. We claim
that all the pretend activities that children engage in, such as
imagining the character, emotions, mental states or even words of
IC, PO or social robots are similar because they involve role play,
more precisely, imagining what it is to be in another’s shoes.
Social robots differ from imaginary companions because they
have an already set appearance, while ICs could take an almost
infinite set of shapes and voices. Moreover, ICs and POs evolve and
grow with children, meaning that their characters and features
change as children develop both emotionally and cognitively. In
other words, ICs grow and develop alongside their creators. But
social robots are not entirely constructed by children, as they
already come embedded with a predefined set of responses and
actions (even if there is a possibility for social robots to adapt their
responses to their users’preferences). This might, in fact, reduce
the chances of children becoming fascinated and absorbed by social
robots, given that sooner or later the cognitive and emotional
limitations of the robots will become apparent to children. What
research on ICs and POs teaches us is that as long as robots are
integrated into children’s play practices, they could very well serve
as instruments that support the development of children’sethical
and dianoetic virtues, just as children’s imaginative practices of
creating imaginary companions do.
Along these lines, we suggest that children-robot friendship,
though not a true (virtue) friendship in the Aristotelian virtue
framework, plays a similar role to children’s moral development
compared to IC and PO and contributes to the development of
ethical and intellectual virtues. In our assessment of children-robot
friendship, we should therefore focus on the role that a social,
relational robot can play. We understand the role of a robot
companion or robot friend as a PRO–Personalised Robotic
Object. We continue with discussing some requirements for
adequateuseofPROsasenablersforchildren’s moral development.
Like a PRO! But Some Restrictions Apply
When children say they want robots to be their friends, this is
actually an invitation: “Let’s play the friendship game.”This play
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184897
Constantinescu et al. Children-Robot Friendship and Virtue Development

is not just aimless entertainment, but an active method of using
imagination, affection, and reason in order to recognize the
perspectives of other (non-human) entities (Davis 2020, 374).
Our claim is that children-robot friendship can play a positive
role in children’s moral development, through the development
of both ethical and intellectual virtues. Furthermore, we suggest
that PROs would have a more direct impact on the acquisition
and practice of intellectual virtues, and a rather indirect impact in
what concerns ethical virtues. Nonetheless, we complement our
optimism over the positive role that PROs might have for
children’s moral development with precaution over
requirements for sound deployment of PROs. We now turn to
discussing PROs as enablers for children’s moral development,
together with the conditions under which PROs might indeed
help children develop as virtuous agents.
First, through their bonding with robot companions, children
exercise the type of dispositions of the soul that enable future
development of ethical virtues. Empirical studies in children-
robot interaction presented in the section on “Children-robot
interaction in practice”show that children consider robots their
friends, they engage in play with these artificial companions, and
even learn from them. In other words, children feel affection
towards robots, although they regard robots as belonging to a
special new ontological category. According to Aristotle, affection
and subsequent care are two natural dispositions of the soul that
must be developed to strengthen a good, virtuous character. Still,
“affection for soulless objects is not called friendship, since the
affection is not mutual”(NE II55b30) at most, it can be called
goodwill (eunoia). Goodwill is a necessary but not sufficient
condition for true friendship, as it is complemented, among
others, by reciprocity. Nonetheless, allowing children to form
affection bonds with robots, although presupposes a non-mutual
type of relationship, can contribute to habituating children with
exercising goodwill and responsibility towards others. This, in its
turn, can indirectly support the development of ethical virtues in
children.
On the cautionary side, we would like to emphasise that it is
important that children (and adults taking care of them)
understand that the type of friendship established with PROs is
not the equivalent of true friendship and that it may not replace
friendship with humans—which means that we need to think of
PROs foremost from the perspective of responsible design. On the
one hand, children should not mistake robot companions for
human companions, and one way to avoid this situation is to
avoid designing robots that “express social cues to deliberately
facilitate bonding with them”(Bartneck et al., 2020, 195). The
seemingly reciprocal affection that robots would display is one
based on modelled calculations and not genuine feelings, which are
constitutive of true friendships. Designing robot companions to
behave as if they were true friends, able to reciprocate virtue-based
affection is misleading, because it gives the impression that the
technology is more advanced than it really is (Riek, 2012). Similar
to the case of IC and PO, it is important to keep explicit the lack of
moral agency in the case of robots, while allowing for one sided
emotional engagement (Bryson, 2018;Mamak, 2021). On the other
hand, the way children interact with robots would need to mimic
their interaction with humans for it to be a good, relevant exercise
for their adult virtuous development: “What is important is that
children interact with robots in certain ways, and conceive of them
in certain ways, that are similar in some respects to how they
interact with and conceive of humans”(Boulicault et al., 2021,7).
Therefore, to fit relevant concerns raised by many in relation to
robots designed with human appearance and functions (Bryson,
2010;Bryson, 2018;Coeckelbergh, 2016), one potential
requirement for PRO design is that the robot-companion could
position itself as a robot in dialogue with the child user, stating
what it can and cannot do, basically stating the limits of the game of
child-robot friendship.
Second, relationships with robot companions may contribute
to the development of dianoetic or intellectual virtues in children,
which are an important acquisition in themselves and are also
essential for the moral growth of a good character. The scientific
and technical nature of robots could add new interesting
challenges to kids’intelligence and contribute to their
reasoning abilities. PROs could thus provide a good stimulus
for studying, experimenting, discovering, tinkering, hacking, for
instance. In particular, PROs may enable knowledge for what is
necessary and what is contingent or of what is invariable and what
is variable (Crisp, 2018, xxiv), contributing to the development of
the intellectual virtue of practical wisdom, which can later guide
them to interpret lived experiences as adults.
An important requirement here is that due attention must be
paid to technical features underlying robot interaction with
children. One condition for building PROs in a responsible way
might involve co-designing them with children (and adults taking
care of them) (Boulicault et al., 2021), hence, to value their
participation and input. Participative co-design has already
proved useful for therapy (see Aslam, van Dijk and Dertien,
2019) and is consistent with the Aristotelian ethical framework,
enabling the exercise of intellectual virtues such as scientific
knowledge or skill knowledge, as well as of ethical virtues such
as perseverance and patience. For PROs to have such a positive
impact on children development, one related requirement might
regard designing the robot-companions as open-source toolkits
that encourage children to experiment, tinker, and imagine, among
others. Available examples include DIT friendly toolkits described
by Aslam et al. (2019) or Vandevelde et al. (2017).
Third, because social robots offer more predictability and less
unclear signals than other humans do, and because children try to
adapt their preferences to the available options (Elder, 2017,
123–24), interpreting robots as friends becomes an exercise in
socialisation and imagination, thus an exercise of their capacities
of seeing certain contexts and relations from multiple points of
view. For children, friendship and play are valuable and go hand
in hand. (Elder, 2017, 121) properly observed that “playful
simulacra can be important in helping children learn skills
that will help them to better navigate the real thing.”The key
is in the relationship’smaking, which happens within the specific
activity of play. To see robots as “artificial”friends is hardly any
different than developing friendships with an imaginary
companion or a personified object: it is a “moral laboratory.”
Related concerns here regard deception and substitution in
children-robot friendship. Nonetheless, worries regarding the
deceptive nature of children-robot interaction, while warranted
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184898
Constantinescu et al. Children-Robot Friendship and Virtue Development

in principle, should not be overplayed. If the wrongness of
deception in robotics is not inherent, but contingent upon the
way in which we assess the implications of children having such
interactions, then in some cases potential deception might not
only be “harmless fun”(Sharkey and Sharkey, 2021, 311), but
even beneficial playfulness (Elder, 2017). Furthermore, while we
acknowledge that due consideration needs to be given to the
potential substitution effect (Sparrow and Sparrow, 2006;Sharkey
and Sharkey 2010), we would like to point out that playing the
friendship game with PROs might complement and enhance
human friendships (Danaher, 2019). As Ryland (2021: 391) puts
it: “it is arguable that befriending robots (to some degree) could
actually improve the extent to which we tolerate and include
other human beings.”A suggestion to conceive and design PROs
with due precaution for deception and potential substitution
effect is therefore to direct the children-robot friendship
towards interaction that enables virtue development on
children’s side. This means taking into account the lack of
moral agency and responsibility in both robot-friends and
child-users and envisioning the role of robots as supporting
virtue development in children’s lives as a whole, beyond their
interaction with robots.
Finally, due precaution needs to be paid to possible harms
resulting from children interacting with PROs and ascribing proper
responsibility and blame. There is yet no definite answer regarding
the network of moral responsibility surrounding robot deployment
and it is not within the scope of the current article to offer one.
Nonetheless, possible candidates for children-robot friendship
resulting in harmful moral outcomes include at least parents or
tutors, designers, programmers, manufacturers, and marketers.
What is important to highlight is that, given the special nature
of the two parties involved in children-robot friendship, with none
being a moral agent, blame, and responsibility for each party is to
be borne by someone else, and this also covers aspects related tothe
amount of time that children spend with their PROs, the functions
that are activated on the PRO relative to children’s group age, or the
way the PRO is empowered to occupy a certain role in the
household, to name but a few.
CONCLUSION
What practical lessons can be derived from this perspective on the
role of PROs in children’s moral development? First, designers
should consider the vastly different types of relationships children
can have, not only with other humans, but also with imaginary
companions and personified objects. Unlike adults, children are
not yet constrained by social norms, thus they have a much bigger
space in which play serves the purpose of discovering, learning,
and imagining new situations and life-contexts. PROs should
thus be co-designed alongside children in order to stimulate
moral imagination and the development of ethical and
dianoetic virtues. Second, PROs should not be marketed for a
specific role or as a substitute to friends, teachers, pets, and so on.
The task of imagining the roles PROs should play in children’s
lives should be left to the beneficiaries of these technologies, that
is children and their tutors. Last but not least, special attention
should be paid to issues related to safety, reliability and
trustworthiness, which we have not addressed in this article,
given our focus on the implications of children-robot friendship
for virtue development. For instance, the use of robot
companions for children raises important questions of privacy
and surveillance, which is an issue for AI embedded technologies
in general. Robots should not become additional tools of
surveillance and control, and this is an issue that should not
be left to be solved by industry self-regulation. Additional
concerns regarding moral and legal responsibility thus arise
when discussing children-robot friendship.
Still, many things remain to be discussed, such as the place to
be ascribed to robots in the established relationship between
children and their parents, teachers, or friends. A co-design
paradigm is undoubtedly recognizing children’s input and
perspectives, but once the PROs come out the factory gate,
should children be allowed to tinker their own robots, to mod
them like a game console, to modify them to suit their own
interests beyond the producers’bounds?
DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in
the article/supplementary material, further inquiries can be
directed to the corresponding author.
AUTHOR CONTRIBUTIONS
All authors listed have made an equally substantial, direct, and
intellectual contribution to the work and approved it for
publication.
FUNDING
This work was supported by a grant of the Romanian Ministry of
Education and Research, CNCS—UEFISCDI, project number
PN-III-P1-1.1-TE-2019-1765, within PNCDI III, awarded for
the research project Collective moral responsibility: from
organizations to artificial systems. Re-assessing the Aristotelian
framework, implemented within CCEA & ICUB, University of
Bucharest (2021–2022).
ACKNOWLEDGMENTS
We would like to thank the editors of the research topic “Moral
Responsibility of Robots”and the article reviewers for their
instructive suggestions. Furthermore, we thank Emilian
Mihailov for insightful comments on earlier drafts of the
paper. We would also like to thank the audience of the
“Responsibility Matters Workshop Series”[21 Oct 2021,
hosted by the Research Center in Applied Ethics (CCEA),
University of Bucharest], for pertinent observations that
helped us to improve our work.
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 8184899
Constantinescu et al. Children-Robot Friendship and Virtue Development

REFERENCES
Ali, S., Mehmood, F., Dancey, D., Ayaz, Y., Khan, M. J., Naseer, N., et al. (2019b).
An Adaptive Multi-Robot Therapy for Improving Joint Attention and Imitation
of ASD Children. IEEE Access 7, 81808–81825. doi:10.1109/ACCESS.2019.
2923678
Ali, S., Moroso, T., and Breazeal, C. (2019a). “Can Children Learn Creativity from a
Social Robot?,”in Proceedings of the 2019 on Creativity and Cognition, San
Diego, CA, USA, June 23-26, 359–368.
Aslam, S., van Dijk, J., and Dertien, E. (2019). “CoCoCo: Co-Designing a Co-design
Toolkit for Co-bots to Empower Autistic Adults,”in Paper presented at 4th
RTD Conference: Design United, Rotterdam, Netherlands, 19-22 March, 1–16.
Bartlett, B., Estivill-Castro, V., and Seymon, S. (2004). “Dogs or Robots—Why Do
Children See Them as Robotic Pets Rather Than Canine Machines?,”,
Auckland, New Zealand, 20 - 23 January. (AUIC2004).Proc. Fifth
Australasian User Interface Conference
Bartneck, C., Belpaeme, T., Eyssel, F., Kanda, T., Keijsers, M., and Sabanovic, S.
(2020). Human-Robot Interaction. An Introduction. Cambridge: Cambridge
University Press.
Belpaeme, T., Baxter, P., de Greeff, J., Kennedy, J., Read, R., Looije, R., et al. (2013).
“Child-Robot Interaction: Perspectives and Challenges,”in Social Robotics.
Editors G. Herrmann, M. J. Pearson, A. Lenz, P. Bremner, A. Spiers, and
U. Leonards (NY: Springer International Publishing), 452–459. doi:10.1007/
978-3-319-02675-6_45
Belpaeme, T., Kennedy, J., Ramachandran, A., Scassellati, B., and Tanaka, F. (2018).
Social Robots for Education: A Review. Sci. Robot. 3 (21), eaat5954. doi:10.1126/
scirobotics.aat5954
Beran, T., and Ramirez-Serrano, A. (2010), Do children Perceive Robots as Alive?
Children’s Attributions of Human Characteristics 5th ACM/IEEE International
Conference on Human-Robot Interaction (HRI), Osaka, Japan. 02-05 March.
137–138. doi:10.1109/HRI.2010.5453226
Blanson-Henkemans, O., Looije, V., Alpay, L., Neerincx, R., Hoondert, L. L., and
Groot, M. A. (2012). &ldquo;I Just Have Diabetes&rdquo;: Children&rsquo;s
Need for Diabetes Self-Management Support and How a Social Robot Can
Accommodate Their Needs. Pi 4, 51–61. doi:10.2147/PI.S30847
Boulicault, M., Phillips-Brown, M., Kory-Westlund, J. M., Nguyen, S., and Breazeal,
C. (2021). Authenticity and Co-design: On Responsibly Creating Relational
Robots for Children. Algorithmic Rights Prot. Child. Available at: https://wip.
mitpress.mit.edu/pub/authenticity-and-co-design/release/1. doi:10.1162/
ba67f642.646d0673
Breazeal, C., Harris, P. L., DeSteno, D., Kory Westlund, J. M., Dickens, L., and
Jeong, S. (2016). Young Children Treat Robots as Informants. Top. Cogn. Sci. 8
(2), 481–491. doi:10.1111/tops.12192
Brink, K. A., and Wellman, H. M. (2020). Robot Teachers for Children? Young
Children Trust Robots Depending on Their Perceived Accuracy and Agency.
Dev. Psychol. 56 (7), 1268–1277. doi:10.1037/dev0000884
Bryson, J. J. (2018). Patiency Is Not a Virtue: The Design of Intelligent Systems and
Systems of Ethics. Ethics Inf. Technol. 20, 15–26. doi:10.1007/s10676-018-
9448-6
Bryson, J. J. (2010). “Robots Should Be Slaves,”in Close Engagements with Artificial
Companions: Key Social, Psychological, Ethical and Design Issues. Editor
Y. Wilks (Amsterdam: John Benjamins Publishing Company), 63–74.
doi:10.1075/nlp.8.11bry
Cappuccio, M. L., Peeters, A., and McDonald, W. (2020). Sympathy for Dolores:
Moral Consideration for Robots Based on Virtue and Recognition. Philos.
Technol. 33, 9–31. doi:10.1007/s13347-019-0341-y
Coeckelbergh, M. (2020). Artificial Intelligence, Responsibility Attribution, and a
Relational Justification of Explainability. Sci. Eng. Ethics 26, 2051–2068. doi:10.
1007/s11948-019-00146-8
Coeckelbergh, M. (2016). “Is it Wrong to Kick a Robot? Towards a Relational and
Critical Robot Ethics and beyond,”in What Social Robots Can and Should Do.
Editors Johanna Seibt, Marco Nørskov, and Søren Schack Andersen
(Amsterdam: IOS Press), 7–8.
Coeckelbergh, M., Pop, C., Simut, R., Peca, A., Pintea, S., David, D., et al. (2016). A
Survey of Expectations About the Role of Robots in Robot-Assisted Therapy for
Children with ASD: Ethical Acceptability, Trust, Sociability, Appearance, and
Attachment. Sci. Eng. Ethics 22 (1), 47–65. doi:10.1007/s11948-015-9649-x
Constantinescu, M., and Crisp, R. (2022). Can Robotic AI Systems Be Virtuous and
Why Does This Matter? Int. J. Soc. Robotics. doi:10.1007/s12369-022-00887-w
Constantinescu, M., Vică, C., Uszkai, R., and Voinea, C. (2022). Blame it on the AI?
on the Moral Responsibility of Artificial Moral Advisors. Philos. Technol. 35. 1.
doi:10.1007/s13347-022-00529-z
Constantinescu, M., Voinea, C., Uszkai, R., and Vică, C. (2021). Understanding
Responsibility in Responsible AI. Dianoetic Virtues and the Hard Problem of
Context. Ethics Inf. Technol. 23, 803–814. doi:10.1007/s10676-021-09616-9
R. Crisp (Editor) (2018). Aristotle: Nicomachean Ethics. Second edition
(Cambridge: Cambridge University Press).
Crisp, R. (2015). A Third Method of Ethics? Philos. Phenomenol. Res. 90, 257–273.
doi:10.1111/j.1933-1592.2012.00607.x
Danaher, J. (2020). Robot Betrayal: a Guide to the Ethics of Robotic Deception.
Ethics Inf. Technol. 22, 117–128. doi:10.1007/s10676-019-09520-3
Danaher, J. (2019). The Philosophical Case for Robot Friendship. J. Posthuman
Stud. 3, 5–24. doi:10.5325/jpoststud.3.1.0005
Darling, K. (2021). The New Breed: What Our History with Animals Reveals about
Our Future with Robots. New York: Henry Holt.
Davis, P. E. (2020). “Imaginary Friends: How Imaginary Minds Mimic Real Life,”
in The Cambridge Handbook of the Imagination. Editor A. Abraham
(Cambridge: Cambridge University Press), 373–389. doi:10.1017/
9781108580298.023
Dawe, J., Sutherland, C., Barco, A., and Broadbent, E. (2019). Can Social Robots
Help Children in Healthcare Contexts? A Scoping Review. bmjpo 3 (1),
e000371. doi:10.1136/bmjpo-2018-000371
Di Dio, C., Manzi, F., Peretti, G., Cangelosi, A., Harris, P. L., Massaro, D., et al.
(2020). Shall I Trust You? from Child-Robot Interaction to Trusting
Relationships. Front. Psychol. 11, 469. doi:10.3389/fpsyg.2020.00469
de Graaf, M. M. A. (2016). Evaluation of Human -Robot Relationships. Int. J. Soc.
Robotics 8 (4), 589–598. doi:10.1007/s12369-016-0368-5
Edwards, A., Edwards, C., Spence, P. R., Harris, C., and Gambino, A. (2016).
Robots in the Classroom: Differences in Students’Perceptions of Credibility
and Learning between "teacher as Robot" and "robot as Teacher". Comput.
Hum. Behav. 65, 627–634. doi:10.1016/j.chb.2016.06.005
Elder, A. (2016). False Friends and False Coinage. SIGCAS Comput. Soc. 45 (3),
248–254. doi:10.1145/2874239.2874274
Elder, A. (2017). “Robot Friends for Autistic Children,”in Robot Ethics 2.0: From
Autonomous Cars to Artificial Intelligence. Editors P. Lin, K. Abney, and
R. Jenkins (Oxford: Oxford University Press), 113–126. doi:10.1093/oso/
9780190652951.003.0008
Fernyhough, C., Watson, A., Bernini, M., Moseley, P., and Alderson-Day, B.
(2019). Imaginary Companions, Inner Speech, and Auditory Verbal
Hallucinations: What Are the Relations? Front. Psychol. 10, 1665. doi:10.
3389/fpsyg.2019.01665
Fletcher, N. M. (2016). Envisioning the Experience of Others: Moral Imagination,
Practical Wisdom and the Scope of Empathy. pie 23 (2), 141–159. doi:10.7202/
1070460ar
Gordon, G., Breazeal, C., and Engel, S. (2015). “Can Children Catch Curiosity from
a Social Robot?,”in Proceedings of the Tenth Annual ACM/IEEE International
Conference on Human-Robot Interaction (HRI ’15) (New York, NY, USA:
Association for Computing Machinery), 91–98. doi:10.1145/2696454.2696469
Gunkel, D. (2018). Robot Rights. Cambridge, MA: MIT Press.
Hakli, R., and Mäkelä, P. (2019). Moral Responsibility of Robots and Hybrid
Agents. Monist 102, 259–275. doi:10.1093/monist/onz009
Healy, M. (2011). Should We Take the Friendships of Children Seriously? J. Moral
Educ. 40 (4), 441–456. doi:10.1080/03057240.2011.618774
Howard, D., and Muntean, I. (2017). “Artificial Moral Cognition: Moral
Functionalism and Autonomous Moral Agency,”in Philosophy and
Computing. Editor T. M. Powers (Cham: Springer), 121–159. doi:10.1007/
978-3-319-61043-6_7
Irwin, T. (1999). Introduction. In Aristotle, Nicomachean Ethics Trans. and ed.
T. Irwin, second edition (Indianapolis: Hackett Publishing Company, Inc.), xiii-
xxviii.
Jecker, N. S. (2020). You’ve Got a Friend in Me: Sociable Robots for Older Adults in
an Age of Global Pandemics. Ethics Inf. Technol. 23, 35–43. doi:10.1007/
s10676-020-09546-y
Kahn, P. H., Jr., Freier, N. G., Friedman, B., Severson, R. L., and Feldman, E. (2004).
“Social and Moral Relationships with Robotic Others?,”in Proceedings of the
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 81848910
Constantinescu et al. Children-Robot Friendship and Virtue Development

13th international workshop on robot and human interactive communication,
Kurashiki, Japan, 22-22 September (Piscataway, NJ: Institute of Electrical and
Electronics Engineers), 545–550.
Kahn, P. H., Kanda, T., Ishiguro, H., Freier, N. G., Severson, R. L., Gill, B. T., et al.
(2012). "Robovie, You’ll Have to Go into the Closet Now": Children’s Social and
Moral Relationships with a Humanoid Robot. Dev. Psychol. 48 (2), 303–314.
doi:10.1037/a0027033
Kahn, P. H., Reichert, A. L., Gary, H. E., Kanda, T., Ishiguro, H., Shen, S., et al.
(2011). “The New Ontological Category Hypothesis in Human-Robot
Interaction,”in 6th ACM/IEEE International Conference on Human-Robot
Interaction (HRI), Lausanne, Switzerland, 08-11 March, 159–160. doi:10.1145/
1957656.1957710
Kanda, T., Hirano, T., Eaton, D., and Ishiguro, H. (2004). Interactive Robots as
Social Partners and Peer Tutors for Children: A Field Trial. Human–Computer
Interact. 19 (1-2), 61–84. doi:10.1080/07370024.2004.9667340
Kim, E. S., Berkovits, L. D., Bernier, E. P., Leyzberg, D., Shic, F., Paul, R., et al.
(2013). Social Robots as Embedded Reinforcers of Social Behavior in Children
with Autism. J. Autism Dev. Disord. 43 (5), 1038–1049. doi:10.1007/s10803-
012-1645-2
Kristjánsson, K. (2020). Aristotelian Character Friendship as a ’Method’of Moral
Education. Stud. Philos. Educ. 39 (4), 349–364. doi:10.1007/s11217-020-
09717-w
Mamak, K. (2021). Whether to Save a Robot or a Human: On the Ethical and Legal
Limits of Protections for Robots. Front. Robot. AI 8, 712427. doi:10.3389/frobt.
2021.712427
Manzi, F., Peretti, G., Di Dio, C., Cangelosi, A., Itakura, S., Kanda, T., et al. (2020).
A Robot Is Not Worth Another: Exploring Children’s Mental State Attribution
to Different Humanoid Robots. Front. Psychol. 11, 2011. doi:10.3389/fpsyg.
2020.02011
Matthias, A. (2015). Robot Lies in Health Care: When Is Deception Morally
Permissible? Kennedy Inst. Ethics J. 25 (2), 169–162. doi:10.1353/ken.2015.0007
Meyer, S. S. (2011). Aristotle on Moral Responsibility: Character and cause. Second
edition. Oxford: Oxford University Press.
Michaelis, J. E., and Mutlu, B. (2018). Reading Socially: Transforming the In-Home
Reading Experience with a Learning-Companion Robot. Sci. Robot. 3 (21).
doi:10.1126/scirobotics.aat5999
Mureșan, V. (2007). Comentariu la Etica Nicomahică. revised. Second edition.
Bucharest: Humanitas.
Navon, M. (2021). The Virtuous Servant Owner-A Paradigm Whose Time Has
Come (Again). Front. Robot. AI 8, 715849. doi:10.3389/frobt.2021.715849
Nomura, T., Kanda, T., Kidokoro, H., Suehiro, Y., and Yamada, S. (2016). Why Do
Children Abuse Robots? Is 17 (3), 347–369. doi:10.1075/is.17.3.02nom
Nyholm, S., and Frank, L. E. (2017). “From Sex Robots to Love Robots: Is Mutual
Love with a Robot Possible?,”in Robot Sex: Social and Ethical Implications.
Editors J. Danaher and N. McArthur (Cambridge, MA: MIT Press), 219–244.
Nyholm, S. (2020). Humans and Robots: Ethics, Agency, and Anthropomorphism.
London: Rowman and Littlefield International.
Pashevich, E. (2021). Can Communication with Social Robots Influence How
Children Develop Empathy? Best-Evidence Synthesis. AI Soc 37, 579–589.
doi:10.1007/s00146-021-01214-z
Peeters, A., and Haselager, P. (2021). Designing Virtuous Sex Robots. Int J Soc
Robotics 13, 55–66. doi:10.1007/s12369-019-00592-1
Prescott, T. J., and Robillard, J. M. (2021). Are Friends Electric? The Benefits and
Risks of Human-Robot Relationships. iScience 24 (1), 101993–102014. doi:10.
1016/j.isci.2020.101993
Riek, L. (2012). Wizard of Oz Studies in HRI: A Systematic Review and New
Reporting Guidelines. Jhri 1 (1), 119–136. doi:10.5898/jhri.1.1.riek
Ryland, H. (2021). It's Friendship, Jim, But Not as We Know It: A Degrees-of-
Friendship View of Human -Robot Friendships. Minds and Machines 31,
377–393. doi:10.1007/s11023-021-09560-z
Scassellati, B., Boccanfuso, L., Huang, C.-M., Mademtzi, M., Qin, M., Salomons, N.,
et al. (2018). Improving Social Skills in Children with ASD Using a Long-Term,
In-Home Social Robot. Sci. Robot. 3 (21). doi:10.1126/scirobotics.aat7544
Scassellati, B. (2007). “How Social Robots Will Help Us to Diagnose, Treat, and
Understand Autism,”in Robotics Research. Editors S. Thrun, R. Brooks, and
H. Durrant-Whyte (Berlin: Springer), 552–563. doi:10.1007/978-3-540-
48113-3_47
Severson, R. L., and Carlson, S. M. (2010). Behaving as or Behaving as if?
Children’sConceptionsofPersonified Robots and the Emergence of a New
Ontological Category. Neural Netw. 23 (8-9), 1099–1103. doi:10.1016/j.
neunet.2010.08.014
Sharkey, A. J. C. (2016). Should We Welcome Robot Teachers? Ethics Inf. Technol. 18
(4), 283–297. doi:10.1007/s10676-016-9387-z
Sharkey, A., and Sharkey, N. (2021). We Need to Talk about Deception in Social
Robotics!. Ethics Inf. Technol. 23, 309–316. doi:10.1007/s10676-020-09573-9
Sharkey, N., and Sharkey, A. (2010). The Crying Shame of Robot Nannies. Interact.
Stud. 11 (2), 161–190. doi:10.1075/is.11.2.01sha
Sison, A. J. G., and Ferrero, I. (2015). How Different Is Neo-Aristotelian Virtue from
Positive Organizational Virtuousness? Bus. Ethics A Eur. Rev. 24, 78–98. doi:10.
1111/beer.12099
Sison, A. J. G., and Redín, D. M. (2021). A Neo-Aristotelian Perspective on the Need for
Artificial Moral Agents (AMAs). AI Soc. doi:10.1007/s00146-021-01283-0
Sparrow, R. (2016, Kicking a Robot Dog. 2016 11th ACM/IEEE International
Conference on Human-Robot Interaction (HRI), Christchurch, New Zealand.
07-10 March. 229–229. doi:10.1109/HRI.2016.7451756
Sparrow, R., and Sparrow, L. (2006). In the Hands of Machines? The Future of Aged
Care. Minds Mach. 16, 141–161. doi:10.1007/s11023-006-9030-6
Sparrow, R. (2002). The March of the Robot Dogs. Ethics Inf. Technol. 4, 305–318.
doi:10.1023/a:1021386708994
Sparrow, R. (2021). Virtue and Vice in Our Relationships with Robots: Is There an
Asymmetry and How Might it Be Explained? Int J Soc Robotics 13 (1), 23–29. doi:10.
1007/s12369-020-00631-2
Taylor,M.,Carlson,S.M.,Maring,B.L.,Gerow,L.,andCharley,C.M.(2004).The
Characteristics and Correlates of Fantasy in School-Age Children: Imaginary
Companions, Impersonation, and Social Understanding. Dev. Psychol. 40 (6),
1173–1187. doi:10.1037/0012-1649.40.6.1173
Turkle,S.(2017).Alone Together: Why We Expect More from Technology and Less from
Each Other. New York: Basic Books.
Vallor, S. (2015). Moral Deskilling and Upskilling in a New Machine Age: Reflections on
the Ambiguous Future of Character. Philos. Technol. 28, 107–124. doi:10.1007/
s13347-014-0156-9
Vallor,S.(2016).Technology and the Virtues. A Philosophical Guide to a Future Worth
Wanting. Oxford: Oxford University Press.
Vandevelde, C., Wyffels, F., Vanderborght, B., and Saldien, J. (2017). DIY Design for
Social Robots Available from: https://www.cesarvandevelde.com/static/pdf/diy-
design-for-social-robots.pdf.
Verner,I.M.,Polishuk,A.,andKrayner,N.(2016).ScienceClasswithRoboThespian:
Using a Robot Teacher to Make Science Fun and Engage Students. IEEE Robot.
Autom. Mag. 23 (2), 74–80. doi:10.1109/MRA.2016.2515018
Wallach, W., and Allen, C. (2008). Moral Machines: Teaching Robots Right from Wrong.
Oxford: Oxford University Press.
Werhane, P. (1999). Moral Imagination and Management Decision Making.NewYork:
Oxford University Press.
Zhang, Y., Song, W., Tan, Z., Zhu, H., Wang, Y., Lam, C. M., et al. (2019). Could
Social Robots Facilitate Children with Autism Spectrum Disorders in Learning
Distrust and Deception? Comput. Hum. Behav. 98, 140–149. doi:10.1016/j.chb.
2019.04.008
Conflict of Interest: The authors declare that the research was conducted in the
absence of any commercial or financial relationships that could be construed as a
potential conflict of interest.
Publisher’s Note: All claims expressed in this article are solely those of the authors
and do not necessarily represent those of their affiliated organizations, or those of
the publisher, the editors and the reviewers. Any product that may be evaluated in
this article, or claim that may be made by its manufacturer, is not guaranteed or
endorsed by the publisher.
Copyright © 2022 Constantinescu, Uszkai, Vicăand Voinea. This is an open-access
article distributed under the terms of the Creative Co mmons Attribution License (CC BY).
The use, distribution or reproduction in other forums is permitted, provided the original
author(s) and the copyright owner(s) are credited and that the original publication in this
journal is cited, in accordance with accepted academic practice. No use, distribution or
reproduction is permitted which does not comply with these terms.
Frontiers in Robotics and AI | www.frontiersin.org August 2022 | Volume 9 | Article 81848911
Constantinescu et al. Children-Robot Friendship and Virtue Development