10 NOVEMBER 2017 • VOL 358 ISSUE 6364 729 SCIENCE sciencemag.org
In Living with Robots, Paul Dumouchel and Luisa Damiano foresee a social space inhabited by a variety of artificial agents possessing a spectrum of cogni- tive architectures. These range from entities completely deprived of physical
dimensions, such as virtual banking advisers or military assistants, to social robots
able to stand in for humans in such contexts
as elderly care or special education.
Originally published in France in 2016,
Living with Robots combines the authors’
expertise in philosophy—in particular, Dumouchel’s scholarship on the role of emotion in shaping social life and Damiano’s
work on human and artificial cognition—
to offer insight into problems raised by
advances in robotics and artificial intelligence that will be faced by future societies.
Throughout the book, the authors provide
a conceptual framework for thinking about
possible scenarios of human-robot interactions, most extensively with regard to our
relationships with social robots.
The distinguishing characteristic of social
robots is their ability to engage in emotional,
affective, and empathic exchanges with hu-
By Paula Quinon
A pair of philosophers probe the ethical implications of
designing social robots
The reviewer is at the Department of Philosophy, Lund
University, Lund, Sweden. Email: email@example.com
mans. Because emotional interactions can
leave participants vulnerable to manipula-
tion or abuse, it is essential that we define
ethical guidelines for human-robot interac-
tions. “Synthetic ethics,” the authors argue,
cannot consist of a preconceived set of rules
like the ethical code programmed into au-
tonomous missile defense systems. How we
actually choose to live with social
robots, the types of emotional
exchanges we have with them,
and the progress and direction
in which social robotics develops
will all inform how such guide-
lines take shape and evolve.
Because ethical guidelines
will be shaped by individual exchanges between autonomous
social actors, human engagement
in the design process is twofold.
Artificial empathy will depend
on whether the human designer
chooses to prioritize the superficial display
of humanlike emotions (like “Geminoid,”
an artificial doppelganger built by the ro-boticist Hiroshi Ishiguro to serve as a synthetic stand-in) or whether the robot will
be equipped with an emotional architecture
(like the Massachusetts Institute of Technology’s “Kismet,” a cartoonlike head that
expresses emotions in an “affective loop” created during exchanges with humans).
Less obviously, but most importantly, human actors must be aware of the emotions
that they project onto robots. These emotions
will affect the stereotypes that are perpetuated into all future human-robot interactions.
In Western societies, a generalized anxiety about artificial intelligence pervades
popular culture. This techno-dread may
arise from a larger concern: the fear of the
automatized society that has been present
since the Industrial Revolution. After all, the
word “robot” was first used by Karel Čapek
in reference to a group of humans enslaved
in the role of servants. Revising negative
stereotypes will, according to the authors,
lower the level of anxiety that is associated
with robots and with artificial intelligence.
Fortunately, Western culture
does not have a monopoly on
evaluating human-robot interactions. Dumouchel and Damiano
provide anecdotes about techno-optimism among the Japanese,
many of whom believe that robots
can be friendly, helpful, and even
morally superior to us. Stepping
outside cultural relativism enables us to reconsider the source
of our fear. Instead of imagining
robots stealing our jobs, for example, we are invited to consider
the unexpected success of Paro, a soft robot
designed to mimic a baby seal that—despite
its limited interactivity—has been shown to
have positive effects on the mental health of
people in elder-care facilities.
Living with Robots will meet various expectations, uniting the intellectual depth of
a carefully documented academic treatise
with the pleasure of a casual page-turner.
Those in search of cultural erudition are
provided with myriad references to books
and movies, and those with a taste for technical novelty are treated to fascinating descriptions of the most hi-tech social robots.
According to the traditional account of
cognition, defended by such philosophers
as Descartes, human cognition is the only
type of cognition. This traditional view
needs to be replaced. Humans and robots
have different types of cognitive architectures, true, but neither is ultimately better
or worse than the other.
The future of robotics, reveal Dumouchel
and Damiano, is deeply humanistic, the
ethics of which will emerge from real-life
interactions between human and artificial
agents. As Living with Robots reminds us,
however, we, as the designers and social actors, hold the greater responsibility for the
direction and quality that these relationships ultimately take. j
10.1126/science.aao3113 An elderly woman interacts with a robotic companion at the Leo Polak House in Amsterdam.
Living with Robots
Paul Dumouchel and
Press, 2017. 280 pp.