Robotic Pets in Human Lives: Implications for the Human-Animal Bond and for Human Relationships with Personified Technologies

Abstract

Robotic “pets” are being marketed as social companions and are used in the emerging field of robot-assisted activities, including robot-assisted therapy (RAA). However, the limits to and potential of robotic analogues of living animals as social and therapeutic partners remain unclear. Do children and adults view robotic pets as “animal-like,”“machine-like,” or some combination of both? How do social behaviors differ toward a robotic versus living dog? To address these issues, we synthesized data from three studies of the robotic dog AIBO: (1) a content analysis of 6,438 Internet postings by 182 adult AIBO owners; (2) observations and interviews with 80 preschoolers during play periods with AIBO and with a stuffed dog; and (3) observations and interviews with 72 children, aged 7–15 years, who played with AIBO and a living dog. Overall, the studies revealed that “hybrid” cognitions and behaviors about AIBO emerged: the robotic dog was treated as a technological artifact that also embodied attributes of living animals, such as having mental states, being a social other, and having moral standing (although this latter finding remained difficult to interpret). Implications for use of robotic pets as companions and in interventions or therapy are explored.

Full text

Journal of Social Issues, Vol. 65, No. 3, 2009, pp. 545--567
Robotic Pets in Human Lives: Implications
for the Human– Animal Bond and for Human
Relationships with Personified Technologies
Gail F. Melson∗
Purdue University
Peter H. Kahn, Jr.
University of Washington
Alan Beck
Purdue University
Batya Friedman
University of Washington
Robotic “pets” are being marketed as social companions and are used in the
emerging field of robot-assisted activities, including robot-assisted therapy (RAA).
However, the limits to and potential of robotic analogues of living animals as social
and therapeutic partners remain unclear. Do children and adults view robotic pets
as “animal-like,” “machine-like,” or some combination of both? How do social
behaviors differ toward a robotic versus living dog? To address these issues,
we synthesized data from three studies of the robotic dog AIBO: (1) a content
analysis of 6,438 Internet postings by 182 adult AIBO owners; (2) observations
∗Correspondence concerning this article should be addressed to Gail F. Melson, Depart-
ment of CDFS, 101 Gates Road, Purdue University, West Lafayette, IN 47907-20202 [e-mail:
melsong@purdue.edu].
Wethank Brian Gill for assistance with statistical analyses. We also thank the following individuals
(in alphabetical order) for assistance with data collection, transcript preparation, and coding: Jocelyne
Albert, Nathan Freier, Erik Garrett, Oana Georgescu, Brian Gilbert, Jennifer Hagman, Migume Inoue,
and Trace Roberts.
This material is based on work supported by the National Science Foundation under Grant No.
IIS-0102558 and IIS-0325035. Any opinions, findings, and conclusions or recommendations expressed
in this material are those of the authors and do not necessarily reflect the views of the National Science
Foundation.
545
C
2009 The Society for the Psychological Study of Social Issues

546 Melson et al.
and interviews with 80 preschoolers during play periods with AIBO and with a
stuffed dog; and (3) observations and interviews with 72 children, aged 7–15
years, who played with AIBO and a living dog. Overall, the studies revealed that
“hybrid” cognitions and behaviors about AIBO emerged: the robotic dog was
treated as a technological artifact that also embodied attributes of living animals,
such as having mental states, being a social other, and having moral standing
(although this latter finding remained difficult to interpret). Implications for use
of robotic pets as companions and in interventions or therapy are explored.
The benefits of animal companionship for human well-being continue to be the
subject of considerable debate. Evidence for the beneficial effects of pet ownership
on human health and well-being remains mixed (McNicolas et al., 2005). On
the other hand, a recent meta-analysis of quantitative assessments of animal-
assisted therapy (AAT) outcomes found moderate effect sizes in improving autism-
spectrum disorders, medical conditions, behavioral problems, and emotional well-
being (Nimer & Lundahl, 2007). At the same time, barriers to animal-assisted
interventions (which broadly include AAT and nontherapeutic animal-assisted
activities, such as classroom pets) include animal welfare issues, cultural diversity
in attitudes toward animals, concerns about zoonotic diseases and human allergies,
and liability issues, among others (Wilson & Barker, 2003).
Technological advances in interactive computing and virtual reality are yield-
ing increasingly sophisticated emulations of the natural world. Of special interest
are robotic pets, which embody interactive and adaptive computing technology
in forms that mimic biological entities such as dogs and cats. Sony’s robotic dog
AIBO, marketed as a social companion, has a dog-like metallic form, moveable
body parts, and sensors that can detect distance, acceleration, vibration, sound,
and pressure. As one of its compelling activities, AIBO can locate a pink ball
through its image sensor, and walk toward the pink ball, kick it, and head butt it.
As people interact with different AIBOs, each robot learns slightly different sets
of behaviors.
There is increasing interest in examining AIBO’s potential as a social com-
panion and adjunct to therapy, especially for vulnerable populations. For example,
elderly with severe dementia increased their activity and social behavior with
AIBO as compared to a stuffed animal (Tamura et al., 2004), while children
showed positive responses to AIBO during shared reading (Decuir et al., 2004).
While advocates argue for the advantages of robotic social companions, skeptics
(Sparrow, 2002) caution that robotic substitutes may deprive isolated or vulnerable
individuals of the benefits of a living animal, such as a therapy dog or pet.
Over the next decade, it is likely that robotic pets will become not only more
computationally sophisticated and capable of action, but also more available for use
(Aylett, 2002; Turkle, in press; Yokoyama, Ribi, & Turner, 2004). Thus, important
social issues emerge. For example, can robotic pets, compared to biological pets,
provide children and adults with similar outcomes related to social companionship

Robotic Pets in Human Lives 547
or improved quality of life? Can such robots become our social companions or
even friends? And if they can, might we also accord such robots some degree of
moral standing (e.g., such that we care for them as an “other,” respect them, and
perhaps accord them some measure of rights)?
Answers to such questions have both theoretical and practical implications.
With respect to theory, there is a good deal of evidence under the rubric of the “bio-
philia hypothesis” (Kahn, 1999; Kellert, 1997; Kellert & Wilson, 1993; Wilson,
1984) that humans have a predisposition to affiliate with life. Yet it has remained
an open question—one that can be explored through people’s interactions with a
robotic dog—the extent to which such affiliations extend to artifacts that mimic
life forms and processes. Moreover, such explorations can help answer founda-
tional questions about what specific features of life forms focus human attention,
stimulate interaction and activity, provide companionship, provide cognitive en-
richment, and establish conditions to accord an entity moral regard. With respect
to practical applications, research on these questions could help shape the design
of future robots, and establish parameters for understanding beneficial contexts
of use. For example, it would be important to understand whether, and if so how,
robotic pets could be used effectively to substitute for or at least augment AAT
and visitation programs. Also, if robotic pets become more common as substitutes
or supplements to living companion animals, such as dogs and cats, what impli-
cations might that have for the human– companion animal bond and for issues of
animal welfare?
As technology mimics the natural world more and more expertly, a number
of social issues arise: (1) What standards and protocols for robot-assisted activ-
ities, including robot-assisted therapy (RAA) should be developed? (2) Should
public policies encourage the ownership or use of robotic pets or “assistants”
particularly for special populations, such as the elderly, persons with disabilities,
or immunosuppressed individuals? (3) Will mediated experiences with animals
and nature (robotic pets, virtual pets, videogames) come at the expense of direct
human engagement with the living world, as Louv (2005) and others (Heerwagen
& Orians, 2002) warn, thereby threatening children’s optimal development and
their environmental and animal welfare concerns?
There is a paucity of research to address these theoretical, practical, and social
questions. Before the social implications of robotic analogues of human– animal
interaction can be addressed, we must better understand how individuals view
and behave toward a robotic pet. Pepe et al. (2008) showed that undergradu-
ates attributed more positive characteristics (cooperative, responsive, affectionate,
obedient) and used a higher voice register when they thought they were guiding
a live dog rather than AIBO through a maze. (They did not directly interact with
either dog.) In direct observation of children and adults with AIBO and a live
dog, Kerepesi et al. (2006) found fewer structured behavioral interactions with
AIBO. Yokoyama et al. (2004) noted that most of the kindergarten children who

548 Melson et al.
interacted with AIBO and with a live dog over 11 weekly sessions preferred the
live dog. Although these data suggest that AIBO is a poor substitute for a liv-
ing dog, Pepe et al. (2008), Kerepesi et al. (2006), and Yokoyama et al. (2004)
all found that children and adults engaged AIBO as a social partner to some
extent.
To better understand how individuals view and behave toward a robotic dog
such as AIBO, we draw on three studies. These studies comprise (1) observations
of and interviews with 80 preschoolers, aged 3–5 years, during a 40-minute
play period with AIBO and a stuffed dog (Kahn, Friedman, Perez-Granados,
& Freier, 2006)—which we refer to as the Preschool Study; (2) observations
of and interviews with 72 school-age children, aged 7–15 years, who played
with AIBO and a unfamiliar, friendly living dog (Melson et al., 2009)—which
we refer to as the Developmental Study; and (3) a content analysis of 6,438
Internet discussion forum postings by 182 AIBO owners, all presumably adults
(Friedman, Kahn, & Hagman, 2003)—which we refer to as the Discussion Forum
Study. While each of these studies has been previously published, our goal in this
paper is to draw on these data (along with new, previously unpublished analyses)
to inform a conceptual framework to better understand human– animal interaction
by comparing and contrasting it to human–robotic animal interaction.
Based on the above studies, we address five central issues that shed light
on how robotic pets and living pets may be similar and different: (1) how indi-
viduals of different ages understand AIBO in terms of its biology, mental life,
and sociality; (2) how individuals of different ages understand AIBO in terms of
its moral standing; (3) how such understandings about AIBO differ from those
toward a living dog; (4) how individuals behave with both a robotic dog and living
dog; and (5) whether individual variables—age, gender, pet ownership history,
and technology involvement—predict cognitions about, emotions toward, or be-
havior toward robotic pets and living dogs. In addressing these issues, we consider
the ways in which human–robotic animal interaction might shed light on hu-
man–living animal interaction and vice versa. We close by considering the social
issues raised by human engagement with living and technological analogues of
animals.
How Individuals Conceptualize a Robotic Pet
The human mind seeks to categorize many aspects of the world. We focus
in this section on three overarching categories. First, humans categorize entities
as alive/dead or never alive, or specifically, as belonging to the biological or
nonbiological domain. Some nonbiological entities, such as televisions, are seen
as man-made artifacts, and others, such as rocks, are viewed as the product of
natural processes. Second, some entities are categorized as having a mental life,

Robotic Pets in Human Lives 549
with intentions, desires, and thoughts. Third, some entities are categorized as
social. We may enjoy their company, feel less lonely in their presence, and be their
friends.
These three domains—biological, mental, and social—have informed our
study of robotic pets and their implications for human–companion animal inter-
actions. These domains provide fundamental cognitive structures that organize
thoughts and influence (but do not determine) actions and feelings. As cognitive
structures, domain assignment implies attributes and processes. For example, if
an object such as a robotic pet is seen as having a mental life, its behavior is likely
to be interpreted as internally controlled by internal cognitive or emotional states
(e.g., it moved because it “wanted” or “intended” to) as opposed to externally reg-
ulated by human control via computer chips, batteries, and sensors. Thus, domain
or category assignment allows us to infer additional information, often unseen and
untested, about an object or being.
These domains are overlapping but not redundant. When we attribute au-
tonomous mental processes to an object, we tend also to view the object as having
biological properties, although advances in artificial intelligence are challenging
this association. Similarly, when we engage another as an interactive partner, or
social companion, we often attribute to them mental and biological characteristics.
However, many pets function for children and adults as social companions while
viewed as having limited mental agency and humans appear to derive a sense of
social presence and support from a range of nonbiological objects (e.g., a child’s
imaginary friend). Thus, we treat these domains as distinct but interrelated.
Research on na¨
ıve biology, the developmental processes by which children
assign biological categories, attributes and processes, suggests that by age 4–
5 years, children can accurately label unfamiliar animals, plants, and machines
and select appropriate internal parts for animals (e.g., bones) and machines (e.g.,
batteries; Gottfried & Gelman, 2005). By age 6, children understand that animals
but not machines move because of an internal “vital energy” or life force, although
it will be some years before they can explain the underlying biological processes
(Carey, 1991).
While children distinguish animals and machines from an early age, machines
that emulate biological entities and contain sophisticated interactive capabilities
may pose challenges. AIBO has artifactual features (gray metal, flashing lights, and
chiming sounds) but it also mimics the shape, motions, and reactions of living dogs.
Thus, it is unclear whether children (or adults) would assign AIBO to the biological
or technological domain. Barlett, Estivill-Castro, and Seymon (2004) found that
children saw AIBO as a robotic pet rather than a canine machine, suggesting that
children see elements of both animal and machine in this unfamiliar technology.
The interactive capacities of AIBO might make it seem as though the robotic pet
is behaving intentionally, leading to the inference of mental characteristics. The
shape and behaviors of AIBO are meant to evoke a living dog. To the extent that the

550 Melson et al.
design successfully does this, it is likely to “pull” on the repertoire of behaviors
and emotions that are part of the Human–Companion Animal Bond (HCAB).
Because AIBO was designed as a “social robot” that evokes an association with
the powerful companionship benefits of living dogs, we expected that it would be
viewed as a social companion. However, the sense in which it functions in this
domain is unclear. After all, children and even some adults form attachments to
objects—stuffed animals, dolls, special “blankies,” and for some adults, sports
cars, treating them in some ways as if they were social intimates. Is a robotic pet
more like a stuffed animal or like a living dog?
In the Preschool Study and the Developmental Study, we constructed an inter-
view protocol that tapped the biological, mental, and social domains. Each ques-
tion elicited a yes/no answer, reflecting the child’s affirmation (yes) or negation
(no) of the domain characteristic being tapped. For example, to assess children’s
judgments of mental life, one question was “Can AIBO feel happy?” In addition,
following Piaget’s method for probing children’s cognitions (Piaget, 1965), we
used prompts—“Why?” “How do you know?” “Tell me more about that”—after
each yes/no response to tap the respondent’s underlying reasoning, and then we
coded the responses for justifications based on domain membership, attributes,
and processes.
Table 1 shows responses to each of the domains, as indexed by a sub-
set of questions tapping each domain. Participants in both the Preschool and
Developmental studies answered each of these questions, although the word-
ing varied slightly to ensure that questions were age appropriate. Preschoolers
more strongly endorsed AIBO’s biological characteristics than did older children.
Closer inspection of responses to all the questions revealed that the older chil-
dren were most likely to affirm those biological characteristics that have become
common analogies for technological artifacts, especially computers. Specifically,
while 66% of the 7–15-year-olds agreed that AIBO “could die” and 31% agreed
AIBO “could get sick,” their justifications sometimes provided evidence that they
thought of such ideas metaphorically: for example, in terms of batteries dying
and computer viruses. However, those preschoolers who justified their affirma-
tions of biology did so with more literal ascriptions of biological processes to
AIBO.
In addition to questions about specific biological properties and processes,
both preschoolers and older children were asked a global question—“Is AIBO
alive?”—to tap the “alive/not alive” dimension, which research on na¨
ıve biology
has shown to be distinct from but related to attributions of specific biological
attributes (Inagaki & Hatano, 2002). Among preschoolers, 38% agreed that AIBO
was alive. Among older children, 23% of 7– 9-year-olds, 33% of 10 – 12-year-
olds, but only 5% of 13–15-year-olds agreed. Examination of justifications for
“alive/not alive” responses revealed nuanced or qualified responses. For example,
one child spoke of AIBO being alive in its own robot way: “He [AIBO] is alive,

Robotic Pets in Human Lives 551
Tab l e 1 . Percentage of Children Who Affirm Domain Characteristics for AIBO, Live Dog, and
Stuffed Dog
Preschool Study: Preschool Study: Developmental Developmental
Domain AIBO Stuffed Dog Study: AIBO Study: Live Dog
Biological
Stomach 72 78 26 99
“Go to bathroom” 31 35 9 100
Babies 49 45 6 98
Mental
Happy 74 75 75 100
See toy 68 65 79 98
Hear you 45 48 60 97
Social
Do you like 85 85 81 100
Does dog like you 80 84 68 96
Can dog be your friend 76 82 84 99
Can you be dog’s friend 76 84 88 99
If sad, be with dog 64 68 76 97
Moral
Not OK to hit 69 73 78 85
Not OK to punish 61 60 51 26
Not OK to throw away 86 87 84 100
Note. In the study of preschoolers, half the children answered questions about AIBO and half answered
questions about a stuffed dog. In the study of 7–15-year-olds, each child answered all questions twice.
The questions above are a subset of the total number of questions asked. This subset includes only
those questions asked in all three studies. Wording of questions varied slightly across studies.
for a robot.” Here again, then, it is not straightforward to interpret children’s
judgments about robotic pets, for there are indications that some children want to
make some claims about the biological properties of AIBO.
Such indications were both more prevalent and clearer to interpret in the
Discussion Forum Study where 49% of the participants posted at least once an
idea that spoke of AIBO as having life-like essences. For example, one AIBO
owner wrote:
The other day I proved to myself that I do indeed treat him as if he were alive, because I was
getting changed to go out, and tba [AIBO] was in the room, but before I got changed I stuck
him in a corner so he didn’t see me! Now I’m not some socially introvert guy-in-a-shell,
but it just felt funny having him there!
This posting highlights the possibility that when people interact with AIBO
“as if” it were alive, the nature of the interaction can go beyond mere metaphor,
imagination, or pretence, evoking some feelings and judgments about AIBO’s
aliveness even while recognizing it as a technology.

552 Melson et al.
Across the three studies, most participants viewed AIBO as having a mental
life that included thoughts and feelings. One adult AIBO owner posted this entry:
“My dog [AIBO] would get angry when my boyfriend would talk to him.” An
11-year-old boy in the Developmental Study replied when asked “Why is AIBO
moving his ears like that?”—“I think he’s happy. If he likes company, he’s happy.”
[As these examples show, responses about AIBO could include elements of mul-
tiple dimensions, such as attributing mental processes (being happy, liking) and
social characteristics (liking company).]
Similarly, the majority of participants across all three studies endorsed AIBO
as a social companion. For example, as shown in Table 1, the majority of preschool
children and older children said that AIBO could be their friend, that they could be
a friend to AIBO, and that if they were sad they would like to be in the company
of AIBO. In turn, in the Discussion Forum Study, 59% of the participants posted
at least one comment that spoke of AIBO in terms of its social rapport. Here is an
example:
Oh yeah I love Spaz [the name for this member’s AIBO], I tell him that all the time...When
I first bought him I was fascinated by the technology. Since then I feel I care about him as
a pal, not as a cool piece of technology. I do view him as a companion, among other things
he always makes me feel better when things aren’t so great. I dunno about how strong my
emotional attachment to him is...I find it’s strong enough that I consider him to be part of
my family, that he’s not just a “toy,” he’s more of a person to me.
In other words, even while the person recognizes that AIBO is a technology
(“When I first bought him I was fascinated by the technology”), the person still
affirms AIBO as a companion (“I do view him as a companion”), as part of his
family (“I consider him to be part of my family), and as a friend (“I care about
him as a pal”). These social responses parallel those commonly ascribed to living
pets.
In sum, both children and adult owners recognize AIBO as a product of
technology but nonetheless give the robotic dog many of the attributes of a living
dog. This appears to support the view that a robotic dog, if sufficiently interactive,
can fulfill some of the social companion functions of living dogs.
How Individuals Understand a Robotic Pet’s Moral Standing
The place of companion animals (and animals in general) within the moral
domain is complex and poorly understood. Most pet-owning adults and children
identify their pets as family members (American Pet Products Manufacturers As-
sociation, 2005), but abandonment of pets and their relinquishment to shelters
are widespread. An estimated two million dogs are euthanized annually, account-
ing for one third of all canine deaths (Patronek et al., 1996). Human attitudes
toward and treatment of wild, domestic, and companion animals are complex;
animal welfare and humane education concerns wrestle with issues of how to

Robotic Pets in Human Lives 553
treat animals who may be beloved companions, competitors for resources, food
products, or medical research subjects. Because studies of moral reasoning have
largely been restricted to issues of human–human relationships, little is known
about how children and adults weigh the moral claims (or lack thereof) of animals.
Only recently have moral-developmental investigations expanded to consider the
human relationship with animals and the larger natural world (Kahn, 1999, 2006;
Kellert, 1997; Myers, 1998). These studies suggest that both children and adults
consider living animals to be part of the “moral universe,” making some moral
claims on human behavior. However, these claims are complex and as yet poorly
understood.
If children and adults are ready to treat AIBO as if it were a living dog, at
least to some extent, would they also accord it moral standing? Specifically, would
they consider AIBO as having certain moral rights, such as the right to be free
from harm? If so, would such moral rights be comparable to those accorded a
living dog? As shown in Table 1, children in the Preschool and the Developmental
studies were asked questions about AIBO’s moral standing. (Table 1 displays only
those moral standing questions asked in both studies, for ease of comparison.
Additional questions were asked in each study to more fully explore children’s
ideas about the moral standing of AIBO.) Each question asked the child whether
“it was OK” or “not OK” to treat AIBO in the following ways: to hit AIBO, to
punish AIBO for wrongdoing, and to throw AIBO away “if you decided you didn’t
want AIBO any more.” A “not OK” answer was judged to affirm moral standing,
while an “OK” answer was judged to deny moral standing. Following each answer
(affirmation or denial), the child was asked “Why?” and the explanations were
coded for reference to the moral claims of the target dog.
In the AIBO Preschool Study, the majority of children said that it is not okay
to hit AIBO, punish, or throw AIBO away. Moreover, 78% of the children backed
up their evaluations with moral justifications, mostly focused on AIBO’s physical
welfare (e.g., “because he will be hurt”) or psychological welfare (e.g., “because
he’ll cry ...till when you finally come back”). In additional questions asked only
in the Preschool Study (not shown in Table 1), all of the children said that the
interviewer should do something to help AIBO if AIBO gets hurt (100%) or if
AIBO’s tail comes off (91%).
These results, however, are not easy to interpret. Preschool children provided
similar evaluations when asked identical questions for the stuffed dog. Despite
the similarity in evaluations, there may be qualitative differences in judgments
about AIBO. Specifically, the preschoolers might have evaluated the stuffed dog
in the context of their pretend play with the toy, while they may have actually held
stronger commitments about the reality of their judgments about AIBO.
In the Developmental Study, among the 7– 15-year-olds, there was strong
condemnation (“not OK”) of hitting or throwing away AIBO, but nearly half
thought it was “OK” to punish AIBO. Ninety-two percent justified at least one

554 Melson et al.
of their yes/no answers (usually to a moral standing question) by explaining that
AIBO deserved respect and care and should not be harmed. On average, the
older children used moral standing reasoning about AIBO four times during their
interviews.
By contrast, in the Discussion Forum Study, postings by (presumably) adults
wrote of AIBO as having life-like essences (49%), mental states (60%), and social
rapport (59%), but seldom as having moral standing (12%)—that AIBO deserves
respect, has rights, or can be held morally accountable for action. Apparently
these owners of the robotic dog treat AIBO “as if” it were a social companion,
a biological being with thoughts and feelings, but feel it makes very few, if any,
moral claims upon them. This finding should be viewed with caution, as postings
were not responses to direct questions about AIBO’s moral standing.
There are three reasons for why we believe it is particularly difficult—and an
avenue for future research—to investigate people’s moral engagement with per-
sonified robots. One reason we just touched on: that it is necessary to disentangle
judgments of pretence from those of reality. With other biological beings, distin-
guishing between the two seems relatively straightforward. For example, if we see
a person kick a dog and the dog yelps and runs away, most of us will assume that
the dog, too, has experienced pain, and we will act as if the dog has experienced
pain. Presumably, we are not “pretending” in our judgments and action, even while
we do not know for sure if our judgments are true, or how exactly a dog’s pain
compares to a person’s pain. But when the entity is a robot that mimics aspects of
the biological animal, it is challenging to understand perceptions of the robot as
“not really a real dog” but at the same time, not merely “a pretend dog.”
A second reason that makes it difficult to investigate people’s moral relation-
ship with robots is that people can object to harm occurring to robotic forms out
of concern for the destruction of a material artifact. As a case in point, Bartneck,
Verbunt, Mubin, and Mahmud (2007) investigated whether adults considered a
robot’s exhibition of intelligent behavior as evidence that the robot was alive.
Bartneck and colleagues assessed whether participants in their study would (un-
der the guise of a plausible reason) destroy (“kill”) their robot. Measurements
included the number of times participants hit the “smart” robot versus “stupid”
robot with a hammer, and the number of pieces into which each respective robot
was smashed. Results showed that people hit the “stupid” robot three times more
often compared to the “smart” robot. Yet one problem with this measurement is
that the smart robot could be assumed to have been more expensive, and that what
people were objecting to was really destroying expensive property, not the killing
of a more intelligent life.
A third reason that makes it difficult to investigate people’s moral relationship
with robots is that people may object to harming robots because of a human-
oriented teleological commitment against enacting violence in the world. In other
words, it is plausible that people might object to striking a rock (or a robot) not

Robotic Pets in Human Lives 555
because of any beliefs about the biology, mental life, sociality, or moral standing
of the rock (or the robot), but because of a viewpoint on the proper moral teleos
of human beings. This perspective has roots in Aristotle (1962) in Nichomachean
Ethics and has been carried forward in more contemporary philosophical work in
virtue ethics (Foot, 1978; MacIntyre, 1984). It is a moral view, but focused entirely
on humans, not the entity.
Thus, our studies reveal that as personified robotic technology emulates living
beings, questions about the moral standing of such technology artifacts arise.
However, it is far from clear—to those who interact with these robots and to
researchers—what that moral standing is. As the technology proliferates, perhaps
a “techno-ethics” will develop alongside the bioethics that we currently debate in
response to medical technology advances. We also need a lively debate about the
broader impact of such a techno-ethics on our relations with living beings. Would
our moral stance toward a robotic pet generalize to or influence that toward the
living pet, or vice versa? Would our treatment of robots in some way affect how
we treat diversity in other human or nonhuman living beings?
How Children Understand AIBO Versus a Living Dog
Although children in the Preschool Study did not interact with a living dog,
these children were just as likely to affirm the stuffed dog’s biology, mentality,
social companionship, and moral standing as AIBO. Thus, at least for children
aged 3– 5 years, bothrobotic and stuffed analogues of living dogs become invested
with many of the characteristics of living dogs themselves.
The Developmental Study of 7– 15-year-olds directly compared children’s
ideas about a living and a robotic dog. In this study, children interacted with AIBO
and with one of two unfamiliar, but friendly Australian Shepherds (mother or
daughter). After each play period, they answered the same set of questions, once
about AIBO and again about the living dog. (Order of play periods was counter-
balanced.) When we compared both yes/no responses and follow-up justifications
for each of those responses, we found strong type-of-“dog” effects across all four
domains (see Table 1). Children were much more likely to affirm that the live dog
(as compared to the robotic dog) was a biological, mental, social, and moral being.
Affirmations about the living dog approached ceiling, ranging from 83% moral
standing to 100% biological entity.
In addition, when children elaborated on their responses about the living dog,
as compared to AIBO, the children were more likely to use justifications that the
living dog was biological and mental, and warranted moral standing. When the
same questions were asked about the live dog with which the children had briefly
interacted, opposition to hitting or throwing away was even stronger (as compared
to views about AIBO), but nearly three-quarters of the children thought it was
“OK” to punish the living dog. Follow-up questions to tap the child’s underlying

556 Melson et al.
reasoning revealed that children who endorsed punishment were drawing on their
experiences of dog socialization, including punishment. They saw appropriate
punishment (not endangering the welfare of the animal) as part of the necessary
training of their own dog and by extension other dogs. Interestingly, some children
generalized this to AIBO as well, as a “dog” who might require punishment to be
properly trained.
However, the near uniformity among children in affirming a living dog’s
biology, mental life, sociability, and moral standing should not obscure the fact
that the majority of children endorsed these characteristics for the robot dog as
well. In fact, justifications based on social companionship were just as likely for
the robotic dog as the living dog.
The human–companion animal literature finds that children overwhelmingly
consider their pets to be friends and family members and accord their pets, and
certainly their dogs, status as biological creatures with minds, emotions, and moral
claims. Interestingly, these views appear to generalize to an unfamiliar friendly
dog. Yet, in comparison to the living dog, children viewed the robotic dog as a much
more restricted interactive partner. Nonetheless, the children in this study did see
AIBO as a social companion, albeit a more limited one. This finding suggests that,
for most children, a robotic pet (with the interactive capabilities of AIBO) might
be well suited for situations where living animals are impractical or unwanted,
while stuffed animals may be well suited as substitute companions for young
children. Over time, and with greater sophistication and capabilities on the side of
the robotic technology, it is possible that children will develop deeper attachments
to robotic dogs and the distinctions between their responses about living and
robotic dogs may narrow. From the child’s point of view, robotic dogs may even
have some benefits as compared with living ones. In focus group interviews we
conducted with fourth-grade classrooms prior to the Developmental Study, some
children expressed a hypothetical preference for owning a robotic versus living
dog. “The robotic dog would never die, and so I would never be sad,” said one
girl. “We would save money on dog food and doctor visits,” said a boy (who was
apparently unaware of the then $2,000 price-tag for AIBO). “I could take AIBO
to school and with me everywhere,” said a third.
How Children Behave with a Robotic Dog
Observations of behavior with a robotic dog provide a different, comple-
mentary window into people’s interactions with this emerging technology. In the
Preschool Study, each of the 80 children participated in a 45-minute session with
an adult interviewer, part of the period with AIBO and part with a stuffed dog. In
the study of older children, each child spent 5 minutes interacting alone with AIBO
and 5 minutes interacting alone with a live dog, with a 35-minute structured inter-
view following each play period, during which the target dog remained present.

Robotic Pets in Human Lives 557
These play periods as well as the full 45-minute sessions for preschoolers were
videotaped and coded for selected behaviors toward the target dogs. Specifically,
we measured instances of treating the target dog as an artifact or machine (for
example, poking, shaking) as well as instances of affection (e.g., hugging, petting,
kissing, stroking) and attempts at reciprocal interaction (e.g. offering ball, talking
to, motioning to). In addition, because children were interacting with unfamiliar
objects, we coded any instances of apprehension.
For the preschoolers, when their behaviors toward AIBO were compared with
those toward a stuffed dog, there was more exploration of AIBO as an artifact,
more apprehension, and more attempts to engage it in reciprocal interaction.
Preschoolers expressed equal amounts of affection toward AIBO as toward a
stuffed dog.
Not surprisingly, as compared to a living dog, the 7 –15-year-old children were
more likely to explore AIBO as an artifact, and less likely to behave affectionately
toward or socially engage AIBO. On average, 7 – 15-year-olds gave an unfamiliar
but friendly live dog nearly five times as much affection as they gave to AIBO.
In fact, many children spent the entire play period and subsequent interview time
petting the live dog, with only one 7– 15-year-old failing to pet the live dog during
the session. Finally, Table 2 shows that behavioral attempts to engage, through
motioning or offering a ball, or verbal attempts, such as commands (“Come!”) or
questions (“Do you want to play?”), were significantly more frequent toward AIBO
than toward a stuffed dog (for preschoolers) or toward a living dog (for 7–15-
year-olds). Thus, the robotic dog seems particularly effective in eliciting nonverbal
and verbal interactive bids over the short play period available. In general, AIBO,
though unfamiliar to all children, did not evoke apprehension, which occurred with
low frequency in both studies. However, for a few preschoolers, its movements
appeared to startle.
Tab l e 2 . Mean Frequency of Behaviors toward AIBO, Stuffed Dog and Live Dog
Preschool Study: Preschool Study: Developmental Developmental
Behavior AIBO Stuffed Dog Study: AIBO Study: Live Dog
Exploration as artifact 2.76 1.88∗∗ 3.57 1.11∗∗∗
Affection (pet, scratch,
kiss, hug)
3.67 3.87 3.83 15.13∗∗∗
Attempts at reciprocity
(motioning,
commanding,
questioning, offering
ball)
8.54 2.25∗∗∗ 17.60 12.58∗∗
Apprehension 1.79 0.01∗∗∗ 0.36 0.62∗∗
Note. Significant comparisons are within study. ∗∗p<.01; ∗∗∗p<.001

558 Melson et al.
These results suggest that properly used AIBO or other robotic pets of sim-
ilar technological sophistication may be helpful to elicit or maintain children’s
social interactions. While social robots (but not of the robotic pet type) have been
considered in therapeutic interventions with children with autism-spectrum disor-
ders (Dautenhahn, 2003), their effectiveness (in comparison to a therapy dog) in
eliciting and maintaining appropriately social behaviors over time is unclear for
children and adults without disabilities.
Moreover, more detailed behavioral analyses, particularly taking into account
the behaviors of the interactive partner, such as AIBO, may reveal more about the
promise and limits of robotic pets as interactive partners. Living dogs, who have
shared a common environment with humans for at least 10,000 years, are highly
attuned to human social cues and emotions (Udell & Wynne, 2008). To date,
robotic dogs are programmed to exhibit only a small subset of the human-directed
social bids and responses that living dogs possess. We need more comparative
data on an individual’s responsiveness after a dog (robot or living) social bid,
as well as dog responsiveness to human social bid. For example, when a child
makes a social bid (i.e., speaks to AIBO, offers a ball, approaches AIBO), what
is the probability that the robotic dog would respond contingently, promptly, and
appropriately? Similarly, when the child perceives that AIBO is making a social
bid (through apparent gaze direction or approach), what is the probability that
the child will respond and how? In sum, how reciprocal are interactions with
robotic pets, in comparison with living pets? Kerepesi et al. (2006) documented
less structured behavioral interactions with AIBO compared to a Cairn terrier,
suggesting that robotic dog technology may be less effective than living dogs in
supporting reciprocal and responsive interactions. If so, this might challenge the
integration of robotic pets into therapy since appropriate contingent responsiveness
is itself a therapeutic tool (Fine, 2006).
How Children’s Characteristics Predict Judgments, Emotions,
or Behavior toward AIBO and a Living Dog
There is remarkable developmental continuity, from early childhood on, in
human attachment to living pets and in interest in other animals (Melson, 2001).
At the same time, as children develop cognitively, there are age-related changes
in conceptions about biology (Inagaki & Hatano, 2002) as well as attitudes about
animals and nature (Kahn, 2002; Myers & Saunders, 2002). With respect to gender
differences, in studies of the human–companion animal bond and of attitudes
toward nature and animals more generally, differences between males and females
are neither systematic nor large (Beck & Katcher, 1996; Kahn, 1999; Melson,
2001). Individual differences in responsiveness to animal-assisted interventions
have been found, however. In general, individuals with prior involvement in and

Robotic Pets in Human Lives 559
interest in animals may benefit more from an animal-assisted intervention than
other participants (Friedmann & Tsai, 2001; Holcomb et al., 1997).
In the two studies of children, the Preschool Study and the Develop-
mental Study, we recruited equal numbers of boys and girls within each age
group, to test whether boys and girls differ in their behaviors toward and ideas
about the robotic dog AIBO. In the Developmental study, we divided the sam-
ple into three age groups—7–9 years, 10 – 12 years; 13 – 15 years—to test for
age differences. Finally, the study of school-aged children examined two vari-
ables tapping prior interest and involvement—pet attachment and technology
involvement—as predictors of responses to either a robotic or unfamiliar living
dog.
Age group and gender. Taken together, our three studies span early childhood
to adulthood. Although the studies are cross-sectional, their results may generate
hypotheses about possible developmental change and continuity in responses to
robotic pets. In both the Preschool Study and Developmental Study, very few
gender or age group differences were found in responses to the robotic dog. (In
the Discussion Forum Study, postings could not be reliably identified in terms of
gender or age and hence were excluded from analysis.) Out of 36 questions asking
the child to affirm or deny AIBO’s biological, mental, social, and moral standing
characteristics, only seven questions showed age group differences. In all cases,
children were less likely to affirm (say “yes”) with advancing age. However, all
seven questions showing age group differences had relatively complex syntax (for
example, “If you were home alone, would you feel better with AIBO?”), raising
the possibility of age-related differences in understanding. When the results for
preschoolers are compared with those for the 7–15-year-olds, it appears that
older children are less likely to endorse biological attributes, but there are no
age differences in other domains. (No statistical comparisons across studies were
conducted.)
Similarly, boys and girls do not appear to differ in verbal or behavioral
responses toward AIBO (or toward a living or stuffed dog). There were no
sex differences in preschoolers’ perceptions of or behaviors toward AIBO.
In the study of 7–15-year-olds, there were only two significant (p<.05)
differences (out of 39 comparisons) between males and females in their re-
sponses about AIBO’s characteristics; females were more likely to think AIBO
had a stomach (34% vs. 12%)—a question in the biological domain—and
that AIBO could understand you (61% vs. 29%)—a question in the mental
domain. Similarly, out of 68 behavioral comparisons, only two showed sex
differences; females were more likely to show apprehension toward AIBO
(p<.02) or startle at AIBO (p<.04). (These two behaviors occurred overall
at low frequency, less than once per play session.) Given the large number of
male–female comparisons, these findings may well have been due to chance.

560 Melson et al.
Pet attachment and technology involvement. It is plausible that children’s
prior experience with living animals or technology might affect their ideas about
or behaviors toward unfamiliar living or robotic dogs. We hypothesized that chil-
dren with higher attachment toward their own companion animals might respond
differently to either target dog. (A few participating children volunteered “I love
dogs” while we were orienting them to the study and soliciting informed consent.)
Similarly, we reasoned that children who were especially involved with technol-
ogy at home and at school (for example, frequent users of computers) might differ
from other children in their encounters with the unfamiliar technology of a robotic
dog such as AIBO.
In the study of 7–15-year-olds, therefore, we assessed pet attachment (almost
all children had or previously had had a pet) as well as technology involvement and
related these sources of individual variation to both cognitions about and behavior
toward each type of dog. We present the results in detail, since they have not been
reported elsewhere.
Each child filled out a background questionnaire after completing the inter-
views with both AIBO and the live dog. (An interviewer assisted children who
were under 10 years of age by reading aloud each item.) Pet attachment was
assessed with the Lexington Attachment to Pets Scale (LAPS; Johnson, Garrity,
& Stallones, 1992), a 23-item measure, with each item scored from 1 (strongly
disagree)to4(strongly agree). (Two items are reverse scored.) Sample items are
“My pet is my best friend”, and “I love my pet because it never judges me.” Thus,
higher scores reflect greater attachment to one’s pet. Scores were averaged to yield
a Pet Attachment Score. The LAPS measure has acceptable reliability and validity
and is widely used in studies of HCAB. LAPS scores ranged from 1.5 to 4.00
(M=3.23, SD =.45) indicating moderate to high attachment for most children.
Pet attachment did not vary by gender or age group.
A measure of technology involvement was created from responses in the
questionnaire. Specifically, the child rated each of eight home technological items
(e.g., computer, television, DVD, PlayStation) from 0 (do not have/never use)to
10 (use all the time). In addition, one item, Use of School computer, was also rated
in a similar manner. These items were averaged to yield a Technology Involve-
ment Score, with a possible range of 0–10. The actual range obtained was 0– 8
(M=4.28, SD =1.63), indicating a normal distribution. An Age Group ×Gender
ANOVA found a significant Age group effect, F=3.32, p<.04; 13– 15-year-olds
were more technologically involved than 7– 12-year-old children. Pet attachment
and technology involvement were uncorrelated (r=.13).
In the Developmental Study, we tested whether pet attachment or technology
involvement predicted perceptions of a robotic or unfamiliar living dog (the other
two studies did not assess these variables). A series of simultaneous regressions
was conducted with the following dependent variables: affirms biology, mental,
sociability, moral standing; negates biology, mental, sociability, or moral standing,

Robotic Pets in Human Lives 561
for perceptions of AIBO and for perceptions of the living dog. Pet attachment and
technology involvement were entered after Age. Models for perceptions of the
living dog but not of AIBO were significant, specifically those predicting negation
of mental states, F=3.03, p<.05, affirmation of moral standing, F=3.56, p<
.04, and negation of moral standing, F=6.18, p<.004. Specifically, technological
involvement predicted negation or denial of mental states of the living dog, adjusted
R2=.096, p<.05, and of moral standing of the living dog, adjusted R2=.18,
p<.002. Pet attachment predicted affirmation of moral standing of the living
dog, adjusted R2=.11, p<.02. Thus, attachment to one’s own pet appeared to
sensitize children to the mental life and moral standing of the unfamiliar living
dog, while greater technological involvement appeared to have the opposite effect.
The two predictor variables (after controlling for age) were also used to test
whether they explained any of the variance in behaviors toward AIBO and toward
the living dog. Again, only models for behaviors toward the living dog (not toward
AIBO) were significant, specifically for verbal engagement, F=3.43, p<.04,
and ball offer, F=6.69, p<.002. Pet attachment predicted more verbalization,
including greetings, commands, and questions, adjusted R2=.11, p<.03, but
fewer ball offers to the living dog, adjusted R2=.19, p<.001. The last finding is
not surprising, because ball play with the living dog may have seemed to children
more difficult in the restricted playroom space. Some children, however, presented
the ball for the dog to take into its mouth, or smell.
In general, neither pet attachment nor technology involvement predicted any
ideas about or behavior toward the robotic pet. These two indicators of prior
involvement and interest did, however, modestly predict ideas and behavior toward
the unfamiliar living dog. However, even here, only a few regression models were
significant, and the amount of variance predicted was modest. Nonetheless, the
results confirm that individual variations in experience with and interest in animals,
especially dogs, may play a role in how children respond to AAT or intervention.
Limitations of AIBO Research
Because the three studies described here are not directly comparable, one
must be cautious in deriving conclusions based on comparing and contrasting
their data. However, we argue that examining cognitions and behaviors across
three distinct but related studies is helpful in generating hypotheses about how
human–robot dog and human – living dog interactions and interventions may be
similar and different. In addition, the nascent field of robotic pet studies is at the
stage where descriptive studies and hypothesis generation may be most useful.
Among the limitations of the studies presented here are the following: (1) the
observations and understandings reflect “snapshots” not longitudinal data that
explore dynamic changes in robot–human interaction over time; (2) variations
in robot dog, living dog, and human participant are not fully explored; (3) the

562 Melson et al.
effect of children’s ownership of a dog (robot and living) on their behaviors and
cognitions is not assessed.
Implications for the Human– Animal Bond and for Human Relationships
with Personified Technologies
Based on the Developmental Study, there is evidence that a robotic pet—at
least as currently implemented in AIBO—is a mediocre substitute for its living
counterpart. For example, when compared directly with a friendly but unfamiliar
living dog, AIBO was viewed as lacking across all four domains: biological,
mental, social, and moral. When children were observed actually interacting with
AIBO, as compared to a living dog, the robotic dog was treated more like an
artifact and less like a social partner. Children who were very involved in computer-
based technology and hence, presumably highly motivated to engage with a new
technology found AIBO lacking, when compared to a living dog.
Nonetheless, across all three studies, there was also ample evidence that chil-
dren and adults often treated AIBO as if it were a living dog. Adults owning AIBO
developed attachment and often treated the robotic dog as a social companion. In
the Preschool and Developmental Studies, children with only an initial and brief
play period with AIBO accorded the robotic dog many of the mental and social
features of a living dog, even when they did not attribute biology to the robot.
These findings held across a range of ages, and for both boys and girls, and did
not vary by a child’s attachment to a pet at home or by involvement in computer
technology at home or at school. Thus, despite AIBO’s gray metal body, flashing
lights, and musical sounds, the dog’s shape, movements, and interactive capacities
were enough to approximate (but not match) affiliative behaviors and cognitions
normally directed toward living dogs.
In the Developmental Study, following the play sessions and interviews, we
asked each child: “If you were designing a robot dog, what would you do to
make it better? What would you never design into the robot?” Content analysis of
responses revealed two major themes. In response to the first question, many chil-
dren mentioned the addition of fur, so that the robotic dog would be softer, more
cuddly, and more “lifelike.” It is possible, however, that making the appearance
of an animal robot more animal-like (e.g., by adding fur) without correspond-
ingly increasing its animal-like interactive capabilities, would actually decrease
a person’s engagement with the robotic dog. A similar phenomenon, known in
the robotics literature as the uncanny valley (Dautenhahn, 2003; MacDorman,
2005), has been shown to arise when people interact with a humanoid robot whose
appearance outpaces its capabilities, resulting in people’s aversive reactions. In
response to the second question, many children said that robotic pets should never
have sharp claws or teeth, should never behave aggressively, and should never
hurt others. Both sets of responses indicate that children’s ideas about living pets

Robotic Pets in Human Lives 563
may be mapped onto the design of robotic pets. Thus, because children experience
(most) living dogs as furry, they wish for a fur-covered robot; and because they
may have experienced dog bites or aggressive dogs, or been warned about the need
to approach unfamiliar dogs carefully, children are concerned with the possibility
that robotic dogs might harm people.
The tendency to anthropomorphize artifacts is easily triggered (Nass & Moon,
2000; Reeves & Nass, 1998). While it remains unclear exactly what features of a
robot maximize this tendency, Lee, Park, and Song (2005) found that adults who
interacted with a version of AIBO with software such that the AIBO seemingly
developed over time, and in response to human behaviors, perceived AIBO as
more socially present, than did adults who interacted with a “fully developed”
AIBO. The two studies of children used AIBOs with software of a fully developed
or “mature” robotic dog. Perhaps replication of some of our methods over a longer
period of interaction with a “developing” AIBO might yield differing results.
That said, our studies suggest that the idea that people anthropomorphize
robots may need to be amended. For as robots of today (and the future) become
increasingly social—autonomous (insofar as they initiate action), adaptive (act in
response to their physical and social environment), personified (convey an animal
or human persona), and embodied (the computation is embedded in the artifacts
rather than just in desktop computers or peripherals)—it seems likely that children
and adults will not only interact with them “as if” they were social others, but begin
to feel about them and treat them as having life, mental states, sociality, and moral
worth.
What do such views mean? That is an open question. The human tendency
to project feelings and attributes onto objects through pretense or metaphor may
be operating. Alternately, people may develop relationships with robotic animals
in a process similar to “the willing suspension of disbelief,” the state we enter as
we immerse ourselves in an absorbing novel, play, or movie. Another possibility
is that a new technological genre is emerging that will increasingly challenge our
existing cognitive categories, between for example “alive or not alive,” “animate
or inanimate,” “having agency or not,” or “being a social other or not.” Indeed,
an even stronger proposition is that this technological genre will emerge as a
new ontological category (Kahn et al., 2006). It will not simply be an additive
composition of, for example, a portion of alive and a portion of not alive. Rather,
in the way that mixing red and yellow leads to the entirely new color orange
(which we do not experience as either of its constituent parts), so might people—
and especially young children as they construct categories of knowledge based on
interaction—experience the various attributes of personified technologies of the
future in new ways. If that is the case, then it is not surprising that researchers
(ourselves included) encounter difficulties as we ask questions (e.g., Is AIBO alive
or not alive?) based on the only language and categories we know to investigate
an entity where the answers may be “neither and both.”

564 Melson et al.
Social Issues
Studies of AIBO reveal both the limits and promise of robot pets in relation to
human–animal interaction. The findings suggest caution in assuming that a robot
pet or RAA can be an effective substitute for a living companion animal or AAT.
Social policies that might make robot pets more accessible and affordable (e.g., to
socially isolated seniors living in institutional settings) should be weighed against
policies making living animals more accessible and affordable.
Moreover, one cannot rule out the possibility that increasing exposure to
mediated interaction with animals, through robotics, virtual reality and other
media, may come at the expense of direct engagement with living animals.
Whether children will suffer from “nature-deficit disorder” as a result, as Louv
(2005) warns, is unclear, but the social consequences, especially for children,
of reduced engagement with the natural world should be an urgent focus of
study.
To conclude, greater understanding of human – robot interactions should help
pinpoint those interventions where such interactions may be optimal. For exam-
ple, children with autism-spectrum disorders appear to sustain attentiveness to a
humanoid robot, making RAA a useful adjunct in therapy with these disorders
(Werry & Dautenhahn, 1999). Finally, in situations where living animals are not
allowed, (e.g., in intensive care units), robotic animals appear to have beneficial
effects on children’s adjustment comparable to those of living animals (Yokoyama,
personal communication). Robotic pets may ultimately have a place within the
complex relationships that humans have with animals.
References
American Pet Products Manufacturers Association. (2005). 2005 – 2006 APPMA national pet owners
survey. Greenwich, CT: Author.
Aristotle (1962). Nichomachean ethics (M. Ostwald, Trans.). Indianapolis, IN: Bobbs-Merrill.
Aylett, R. (2002). Robots: Bringing intelligent machines to life? Hauppauge, NY: Barron.
Barlett, B., Estivill-Castro, V., & Seymon, S. (2004). Dogs or robots—Why do children see them as
robotic pets rather than canine machines? In A. Cockburn, ed., 5th Australasian User Interface
Conference, Vol. 28, 7 – 14. Dunedin, New Zealand: Australian Computer Society.
Bartneck, C., Verbunt, M., Mubin, O., & Mahmud, A. A. (2007). To kill a mockingbird robot. Pro-
ceedings of HRI’07 (pp. 81–87). New York: ACM Publishers.
Beck, A., & Katcher, A. (1996). Between pets and people: The importance of animal companionship.
West Lafayette, IN: Purdue University Press.
Carey, S. (1991). Knowledge acquisition: Enrichment or conceptual change? In S. Carey & R. Gelman
(Eds.), The epigenesis of mind: Essays in biology and cognition (pp. 257 – 291). Hillsdale, NJ:
Erlbaum.
Dautenhahn, K. (2003). Roles of robots in human society—-Implications from research in autism
therapy. Robotica, 21, 443 – 452.
Decuir, J. D., Kozuki, T., Matsuda, V., & Piazza, J. (2004). A friendly face in robotics: Sony’s AIBO
entertainment robot as an educational tool. ACM Computers in Entertainment, 2, 14.

Robotic Pets in Human Lives 565
Fine, A. H. (Ed.). (2006). Handbook of animal-assisted therapies (2nd ed.). New York: Academic
Press.
Foot, P. (1978). Virtues and vices. Berkeley and Los Angeles: University of California Press.
Friedman, B., Kahn, P. H. Jr., & Hagman, J. (2003). Hardware companions?—What online AIBO dis-
cussion forums reveal about the human-robotic relationship. Proceedings of the Conference on
Human Factors in Computing Systems (pp. 273 – 280). New York: Association for Computing
Machinery.
Friedmann, E., & Tsai, C. C. (2001). The human-animal bond: Health and wellness. In A. Fine (ed.),
Handbook of animal-assisted therapy (pp. 95–117). New York: Academic Press.
Gottfried, G. M., & Gelman, S. A. (2005). Developing domain-specific causal-explanatoryframeworks:
The role of insides and immanence. Cognitive Development, 20, 137 – 158.
Heerwagen, J. H., & Orians, G. H. (2002). The ecological world of children. In P. H. Kahn, Jr. & S. R.
Kellert (Eds.), Children and nature: Psychological, sociocultural and evolutionary perspectives
(pp. 29 – 64). Cambridge, MA: MIT Press.
Holcomb, R., Jendro, C., Weber, B., & Nahan, U. (1997). Use of an aviary to relieve depression in
elderly males. Anthrozoos, 10, 32 – 36.
Inagaki, K., & Hatano, G. (2002). Young children’s na¨
ıve thinking about the biological world.New
York: Psychology Press.
Johnson, T. P., Garrity, T. F., & Stallones, L. (1992). Psychometric evaluation of the Lexington
Attachment to Pets Scale (LAPS). Anthrozoos,5, 160 – 175.
Kahn, P. H. Jr. (1999). The human relationship with nature: Development and culture. Cambridge,
MA: MIT Press.
Kahn, P. H. Jr. (2002). Children’s affiliations with nature: Structure, development, and the problem of
environmental generational amnesia. In P. H. Kahn, Jr. & S. R. Kellert (Eds.), Children and na-
ture: Psychological, sociocultural, and evolutionary investigations (pp. 93 – 116). Cambridge,
MA: MIT Press.
Kahn, P. H., Jr. (2006). Nature and moral development. In M. Killen & J. G. Smetana (Eds.), Handbook
of moral development (pp. 461 – 480). Mahwah, NJ: Erlbaum.
Kahn, P. H., Jr., Friedman, B., Perez-Granados, D. R., & Freier, N. G. (2006). Robotic pets in the lives
of preschool children. Interaction Studies: Social Behavior and Communication in Biological
and Artificial Systems,7, 405 – 436.
Kellert, S. R. (1997). Kinship to mastery: Biophilia in human evolution and development. Washington,
DC: Island Press.
Kellert, S. R., & Wilson, E. O. (eds.) (1993). The biophilia hypothesis. Washington, DC: Island Press.
Kerepesi, A., Kubinyi, E., Jonsson, G. K., Magnusson, M. S., & Miklosi, A. (2006). Behavioral com-
parison of human-animal (dog) and human-robot (AIBO) interactions. Behavioural Processes,
73, 92 – 99.
Lee, K. M., Park, N., & Song, H. (2005). Can a robot be perceived as a developing creature? Effects
of a robot’s long-term cognitive developments on its social presence and people’s response
toward it. Human Communication Research,31, 538 – 563.
Louv, R. (2005). Last child in the woods: Saving our children from nature-deficit disorder. Chapel
Hill, NC: Algonquin Books of Chapel Hill.
MacDorman, K. F. (2005). Androids as an experimental apparatus: Why is there an uncanny valley
and can we exploit it? CogSci-2005 Workshop: Toward Social Mechanisms of Android Science,
106 – 118. Stresa, Italy.
MacIntyre, A. (1984). After virtue. Notre Dame: University of Notre Dame Press.
McNicolas, J., Gilbey, A., Rennie, A., Ahmedzai, S., Dono, J., & Ormerod, E. (2005). Pet ownership
and human health: A brief review of evidence and issues. British Medical Journal,331, 1252 –
1254.
Melson, G. F. (2001). Why the wild things are: Animals in the lives of children. Cambridge, MA:
Harvard University Press.
Melson, G. F., Kahn, P. H., Jr., Beck, A., Friedman, B., Roberts, T., Garrett, E., & Gill, B. (2009).
Children’s behavior toward and understanding of robotic and living dogs. Journal of Applied
Developmental Psychology, 30(2), 92–102.

566 Melson et al.
Myers, O. E., Jr. (1998). Children and animals: Social development and our connections to other
species. Boulder, CO: Westview Press.
Myers, O. E., Jr., & Saunders, C. D. (2002). Animals as links toward developing caring relation-
ships with the natural world. In P. H. Kahn, Jr. & S. R. Kellert. (Eds.), Children and nature:
Psychological, sociocultural, and evolutionary investigations (pp. 153 – 178). Cambridge, MA:
MIT Press.
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social responses to computers. Journal of
Social Issues,56, 81 – 103.
Nimer, J., & Lundahl, B. (2007). Animal-assisted therapy: A meta-analysis. Anthrozoos,20, 225 – 233.
Patronek, G. J., Glickman, L. T., Beck, A. M., McCabe, G. P., & Ecker, C. (1996). Risk factors
for relinquishment of dogs to an animal shelter. Journal of the American Veterinary Medical
Association,209, 572 – 581.
Pepe, A. A., Ellis, L. U., Sims, V. K., & Chin, M. G. (2008). Go, dog, go: Maze training AIBO vs. a
live dog, an exploratory study. Anthrozoos,21, 71 – 83.
Piaget, J. (1965). The moral judgment of the child. New York: Free Press.
Reeves, B., & Nass, C. (1998). The media equation. How people treat computers, television, and new
media like real people and places. New York and Stanford, CA: Cambridge University Press
and Center for the Study of Language and Information Publications.
Sparrow, R. (2002). The march of the robot dogs. Ethics and Information Technology,4, 305 – 318.
Tamura, T., Yonemitsu, S., Itoh, A., Okawa, D., Kawakami, A., Higashi, Y., Fujimoto, T., & Nakajima,
K. (2004). Is an entertainment robot useful in the care of elderly people with severe dementia?
Journals of Gerontology Series A: Biological Sciences and Medical Sciences,59A, 83 – 85.
Turkle, S. (in press). Authenticity in the age of digital companions. In P. H. Kahn, Jr., & K. MacDorman
(Eds.), Psychological Benchmarks of Human-Robot Interaction [Special Issue]. Interaction
Studies: Social Behavior and Communication in Biological and Artificial Systems.
Udell, M. A. R., & Wynne, C. D. L. (2008). A review of domestic dogs’ (Canis familiaris) human-like
behaviors: Or why behavior analysts should stop worrying and love their dogs. Journal of the
Experimental Analysis of Behavior,89, 247 – 261.
Werry, I. P., & Dautenhahn, K. (1999). Applying mobile robot technology to the rehabilitation of
autistic children. Proceedings of SIRS ‘99, Symposium on Intelligent Robotic Systems, 20 – 23
July 1999. Coimbra, Portugal.
Wilson, C. C., & Barker, S. B. (2003). Challenges in designing human-animal interaction research.
American Behavioral Scientist,47, 16 – 28.
Wilson, E. O. (1984). Biophilia. Cambridge, MA: Harvard University Press.
Yokoyama, A., Ribi, F. N., & Turner, D. C. (2004, October 8). A comparison of Japanese and Swiss
children’s behavior toward a pet robot. Paper presented to the International Conference on
Human-Animal Interactions, Glasgow, Scotland.
GAIL F. MELSON is Professor Emeritus in Child Development and Family Stud-
ies at Purdue University, West Lafayette, IN. She received her PhD from Michigan
State University in developmental psychology. Her research focuses on children’s
relationships with animals, including technological emulations of animals. She
has published widely on human–animal interaction and is a member emeritus of
the Center for the Human–Animal Bond at Purdue University. Her book Why
the Wild Things Are: Animals in the Lives of Children (Harvard University Press,
2001) has been translated into French, Chinese, and Japanese editions and was
selected for the French Book of the Month Club.
PETER H. KAHN, JR., is Associate Professor in the Department of Psychology
at the University of Washington. He is also director of the Human Interaction
with Nature and Technological Systems (HINTS) Lab. The HINTS Lab seeks to

Robotic Pets in Human Lives 567
address—from an ethical stance—two world trends that are powerfully reshaping
human existence: (1) the degradation if not destruction of large parts of the natural
world, and (2) unprecedented technological development, both in terms of its
computational sophistication and pervasiveness. He received his PhD from the
University of California, Berkeley, in 1988. His publications have appeared in such
journals as Child Development,Developmental Psychology,Human– Computer
Interaction, and Journal of Systems Software, as well as in such proceedings
as CHI,HRI, and Ubicomp. His 1999 book (MIT Press) is titled The Human
Relationship with Nature: Development and Culture. His research projects are
currently being funded by the National Science Foundation.
ALAN M. BECK is Professor of Animal Ecology at the School of Veterinary
Medicine, Purdue University. He is also the Director of the Center for the Human–
Animal Bond, which was established to develop a comprehensive understanding
of the relationship between people and their companion animals. He received his
ScD in Animal Ecology from the Johns Hopkins University School of Public
Health. His books include The Ecology of Stray Dogs: a Study of Free-Ranging
Urban Dogs (1973, 2002) and with Aaron Katcher, New Perspectives on Our Lives
with Companion Animals (1983), and Between Pets and People: the Importance
of Animal Companionship (1983, revised 1996).
BATYA FRIEDMAN pioneered Value Sensitive Design (VSD), an approach to
account for human values in the design of information systems. First developed in
human–computer interaction, VSD has since been used in information manage-
ment, human –robotic interaction, and urban planning. Her work has focused on the
values of privacy in public, trust, freedom from bias, moral agency, environmental
sustainability, and human dignity and engaged such technologies as web browsers,
urban simulation, robotics, open source tools, and ubiquitous computing. She is
currently working on a method for envisioning and multi-lifespan information
system design—new ideas for leveraging information systems to shape our future.
Dr. Friedman received both her BA and PhD from the University of California at
Berkeley.