April
22, 2002 | What is an emotional machine? Usually this
term is applied to a machine-soft- or hardware-that is able to recognize,
express and perhaps even 'have' emotions [1]. The objective of these
machines is to better understand their user and model their behavior
accordingly, in order to provide a smoother and more intuitive interaction.
Think, for instance, of a computer that detects that you are nervous
and actively starts providing tips on how to perform the task at
hand. Or the robot that knows you are tired, and brings you a bottle
of beer as soon as you sit on the couch. (But don't think of the
Microsoft paper clip).
Research in emotional machines resembles, in many ways, research
applied to building intelligent machines. In fact, the two are intimately
related, as it is currently thought that a machine can only be truly
intelligent if it is sensitive to human emotions. One of the main
inspirations for the current boost in this line of research is António
Damasio's famous book Descartes' Error [2] in which he transforms
the famous dictum "I think therefore I am" into "I
feel therefore I am". The basis of consciousness is no longer
exclusively thought, but also emotions.
The goals may have changed, but the methods have not. The procedures
for making an emotional machine are strikingly similar to that of
classic artificial intelligence (AI): First, dissect human emotions,
cut them in their minimal slices. Then, model them into the machine.
This simplistic and individualistic approach is as likely to fall
short here as it did with AI. Human beings are complex beings, in
constant dynamic interaction with their environment and with others,
(re)acting to minimal changes in either. An individual-centered,
dissective approach cannot capture these variables, nor can it capture
the "immediate coping" [3] that accounts for most of our
actions.
The dream of creating intelligent emotional machines certainly
reflects advances in technology, but it also reflects the culture
of their creators. In Western cultures, emotional machines are mainly
being created to better serve us. The ultimate dream, at least since
the industrial revolution, though stories of the "Golem"
point to much further back, is one of human leisure assisted by
working robots.
But the Western way is not the only way of thinking about the issue.
In Japan, for instance, different way of thinking about emotional
robots is prevalent, one that escapes the master/servant relationship,
and is open to different ways of thinking about technological life
forms.
Machico Kusahara, an Assistant Professor of Media Research at Kobe
University, Japan, recently gave a talk on this issue, as part of
the Art Creates Change series sponsored by the Ontario College
of Arts and Design and Critical Media. In her lecture Machico described
what is best summarized as Japan's cultural attitudes towards technology.
They are strikingly different from ours.
Japan is world leader in the areas of research and development
of robots. In 1999 it was home to 55% of all industrial robots in
the world [4] and an even larger percentage of recreational robots.
From Japan came the first robot-pets, be it the bygone Tamagotchi;
or the brand new Aibo, a dog with adaptive behavior. These
inventions coexist with large scale, business-oriented applications,
such as Honda's Asimo&trade, a 4-foot, 95-pound, humanoid
robot; and with a series of robots that defy classification, such
as the healing-robots, robots whose only goal is to be looked at
for therapeutic purposes of relaxation, for instance jellyfish robots
[5].
In fact, the proliferation and acceptance of robots in Japanese
culture is so large that when Sony first released its Aibo, it sold
out so quickly that Sony was flooded with letters begging for more!
The demand was so overwhelming that Sony decided to do some research
into Aibo's target group. Sony found that it was constituted mainly
of two main types of consumers: young men who like new gadgets and/or
who are interested in computers (robots as a way to enjoy science
and technology), and people who genuinely enjoy having a robot as
their pet.
Aibo's proud owners dress up their puppies (although this is not
recommended by Sony) and teach them personalized tricks that help
them develop their own personality. The connection between owners
and their pets is so strong and personal, that "that at one
Aibo get-together, owners were able to distinguish their pets from
other Aibo dogs" [6].
The differences become apparent here: Japanese industry invests
heavily into the recreational/leisure robots that seem to nourish
emotions in their users (rather than trying to create robots that
decipher their users emotional states). These users, in turn, are
open to think of these robotic pets as intelligent and emotional
living beings (rather than considering them mere machines to serve
us).
The robot industry is years away of creating the perfect Jeeves
butler, the servant that cares for its owner. However, the technology
to create a robot that is "merely" a friend is already
in place. Pets like Aibo, or the older Tamagotchi, are good examples
of this. These robots do not strive to understand their owner's
emotional state, although Aibo will "understand" when
its owner is angry and pats him (or it?), but they do have the ability
to create emotions in their owners.
In fact, rather than aiming for absolute perfection, in Japan,
a commonly used strategy is to use failure as a way to increase
the realism of the robot. (This is only possible given its entertainment
oriented goals.) For instance, one famous traditional Japanese automata,
the "Bow and Arrow Boy" (yumihiki doji), a doll that shoots 10 arrows,
is programmed to fail at least once for each set [7]. Aibo is also
programmed to ignore its owner every once and then, giving it an
attitude.
By releasing thousands of 'friendly' robots into the commercial
market, Japanese robotic industry progresses not only by getting
feedback from users, but also because these robots have to get adapted
to a variety of people, situations and environments. The knowledge
learned here can then be applied to the creation of more sophisticated,
business-oriented applications. There is a continuum in the progression
from entertainment to "serious" enterprises. But it does
more than this, it also helps people get acquainted, and sympathetic
towards, different life-forms: The robot as a friend that needs
attention and care.
But, why are robot pets such a mass phenomenon in Japan, whereas
in the West, they are regarded suspiciously?
In Japan, says Machico, robots are deemed considerate and friendly.
They are said to have thoughts and souls. This concurs with Japanese
religious beliefs (Shinto and Buddhism). While in the Christian
view of the world God created only people in its own image, in Japan
it is believed that all things in nature have a spirit, there is
not clear distinction between human beings and other life forms.
As a Japanese saying goes, even a 1 inch worm has a half inch soul.
Once you extend this line, how do you distinguish between life and
nonlife?
Think, for instance, of the old Tamagotchi that died when it wasn't
fed properly or simply when it didn't get enough attention and caring.
The consequences of Tamagotchi death were so serious and emotional
for many owners that cemeteries were created for them. When a Tamagotchi
has this kind of reaction in its owner, and when the boundaries
between humans and others is not clear cut, clearly the "life"
category has been extended to it.
Not all has to do with culture and religion of Japan, the particular
socio-demographics of Japan are also at work here. In Japan's overcrowded
large cities it is mostly forbidden to own pets in apartment buildings.
Having a robot-pet, or a relaxation 'healing-robot', is then the
perfect solution. A whole generation of Japanese youngsters is now
being brought up with artificial rather than real pets, and in years
to come it will be interesting to see how this affects them.
As Japan's population ages the demand for robots also increases.
Many Japanese seniors feel isolated and lonely, and the company
of a robot-pet helps them through the day. In addition, as these
robots are increasingly endowed with communicational capabilities,
they start to fulfill a surveillance/monitoring aspect. Besides
keeping company, they can also alert others if something goes wrong
(for instance, if the pet owner does not talk or move for a certain
period of time).
Defining our machines says as much about the machine as it says
about us. The stress on the hierarchical asymmetry between humans
and machines has more to do with our Western conception of ourselves,
than it has to do with the characteristics of the machine. By placing
a totally different set of hopes and fears in the emotional machines
of our creation, by fomenting very different types of human/machine
interactions, Japan is in effect creating a different type of machine.
The Japanese emotional robots are less geared towards understanding
human emotions, than they are towards creating emotions in humans;
it is considerate and friendly, rather than wanting to conquer the
world; and, it is a friend, not a servant whose sole purpose is
to make our life easier. By extending the blurring the differences
between life and nonlife the robot becomes part of us, rather than
one of them.
References:
[1]
Rosalind W. Picard. (1997). Affective computing. Cambridge,
MA: MIT Press.
[2] António R. Damasio. (1994). Descartes Error: Emotion,
Reason and the Human Brain. New York: G. P. Putnam.
[3] Francisco J. Varela. (1999). Ethical Know-How: Action, Wisdom,
and Cognition. Writing Science Series. Stanford, CA: Stanford
University Press.
[4] < http://www.theglobalist.com/nor/factsheets/2001/05-08-01.shtml>
[5] Japan Information Network. (2001, February 1). Robot Pals: Once
an SF Dream, Now a Reality. <http://jin.jcic.or.jp/trends/article/010204ev_r.html>
[6] David Pogue. (2001, January 25). Looking at Aibo, the Robot
Dog. The New York Times. <http://www.nytimes.com/2001/01/25/technology/25STAT.html>
[7] This automata was created by one of Japan's most famous artisans,
Tanaka Hisashige (1799-1881)
bio:
Ana
Viseu is a researcher currently working at the University
of Toronto on her Ph.D. dissertation entitled, "Socio-technical
worlds: The visions and realities of bodynets", which focuses
on the development and implementation of wearable computers and
the emerging sociotechnical worlds that sustain it. Her research
interests include questions of privacy, social dimensions of technology,
and the mutual adaptation processes between individuals and technology.
Ana is the director of the 'Privacy Lecture Series' in Toronto,
which serves as a forum to foment awareness and discuss different
facets of privacy. Ana holds a Master's Degree in Interactive
Communication from the Universitat AutÚnoma de Barcelona, Spain.
|
advertise
here
email for info
|