10 , 2003
Flesh and Machines: How Robots Will Change Us is the latest
book by Rodney Brooks, Director of the Massachusetts Institute of
Technologys (MIT) Artificial Intelligence Laboratory. The
authors affiliation is telling. It informs the reader of Brooks
academic and research accomplishments and, in addition, it prepares
him/her for the enthusiastic, techno sophisticated view of the world
for which the MIT is well known. Both premises are accurate. Flesh
and Machines is a well researched book on the history and development
of robotics and artificial intelligence (AI). Brooks utilizes a
jargon-free vocabulary to develop his arguments, and illustrates
them with real life examples taken mostly from his personal
experience. Thus, in spite of its fairly esoteric topic, Flesh
and Machines is of easy access to all, even those with no prior
knowledge of the subject.
Unfortunately Flesh and Machines falters where so many other
science and technology books have failed before, in its overly simple
conceptual treatment of technology. Pop high-tech books unbridled
technological visions of the future are often accompanied by scant
treatments of the social transformations that create, and accompany
it. Technologies are neutralized as tools, emptied of content or
context. From this perspective artefacts are produced in the lab,
by engineers and other geeky types, and are then consumed by the
public at large. Period. Flesh and Machines is no exception.
The feeble analysis of the social dimensions that accompany the
high-tech society it postulates is not only worrisome but somewhat
Key to understanding of Flesh and Machines is Brooks
definition of robots. He avoids the salvation or damnation dichotomy,
that presents robots as the path to immortality or towards serfdom,
by adopting a broader than usual definition of robots. Robots are
not only mechanical beings (machine-machines), but also
the entities that result from the merger of humans and machines
(man-machines). In Brooks view, the human elements
will not, so to speak, be lost, but rather augmented thus remaining
always one step ahead of the machine. Flesh and Machines
deals with the evolution of both species.
For the sake of argument Flesh and Machines can be divided
in three major sections. In the first Brooks recounts the history
and evolution of the field of AI and details his contribution to
it. In the second, the author projects the short term future of
robotics, and portrays some of the features of the society that
will embrace them. Finally, in the third and last part, Brooks turns
to the ageless issue of the difference between humans and machines,
and discusses the merger of robots and humans as the third
Brooks contribution to the development and transformation
of AI is noteworthy in more than one way for it challenged (and
continues to do so) many of the sacred and unquestioned principles
of artificial intelligence research.
At a time when many were approaching artificial intelligence through
the modeling and manipulation of symbols following logical rules,
that is, creating complex worlds in which all behaviors are accounted
for, Brooks developed a biological approach to AI. This biological
framework takes as a starting point that intelligent creatureshuman
or notare situated and embodied, that is, they are autonomous,
rather than being controlled by a third party, and exist in an environment,
constantly reacting to it. Thus, cognition does not result from
withdrawing from the task at hand and analyzing it with step by
step, but from direct, lived, immediate experience of the environment.
Intelligence cannot be separated from its lived experience.
Brooks then went a step further, and set out to take cognition
out of its pedestal and replace it with perception and action. What
if, he asked, reasoning is not the basis of cognition at all, but
rather the ability to sense and react to the world as we encounter
it? After all, beings like insects, display vast amounts of intelligent
behavior, greater than most robots, since they can look for food
and hideouts, avoid obstacles, mate, etcetera, but still score fairly
low in the traditional IQ scale.
Using this approach, and a subsumption architecturecreating
layers of behavior that interact and regulate one anotherBrooks
has built, and supervised the creation of, several robot-creatures
which he describes with great detail and minutiae in Flesh and
Brooks approach to robotics and cognition resembles that
of Francisco Varela, the noted cognitive scientist, who also advanced
an embodiedbased on perception and actionapproach to
cognition. Varela proposed that we can engage with the world in
two different ways, we can do it in an analytical, abstract, detached
fashionthat resembles the mandates of classic AI,or
in a spontaneous, immediate way. To the first he gives the name
of know-what, and to the second that of know-how, or immediate coping.
According to Varela, most of our mental and active life is
of the immediate coping variety, which is transparent, stable, and
grounded in our personal history (1999). Hence, he puts forward
the idea of cognition as enaction, that is, the world is not pre-given
or constructed by the subject, but rather interpreted through each
his/hers lived cognition, that is, through each entitys embodied
sensorimotor structures. Leaving the body behind to create intelligent
machines is then simply not possible.
One of the important implications of this approach, not only for
AI but for a general understanding of the world, is that it bridges
the Human versus Others cognitive split. The common trait amongst
all living cognitive beings is that they possess a know-how
constituted on the basis of the concrete (ibid).
So far so good. It is when Brooks starts projecting the future
applications of robots that the argument starts to falter. Brooks
defends that there are currently two major problems with robot development:
batteries and navigation. He then argues that remote presence,
a sort of wireless (super)vision, is the most likely short term
solution to this problem. Imagine, for instance, the case of the
ordinary individual who is sitting in a taxi, rushing
towards the airport and suddenly realizes that he/she left the stove
on. Going back to check it is impossible since the flight would
be lost. Leaving it on is equally out of the question since the
house may burn down. If instead of having to go back home this person
could simply robot-in and guide the robot towards the
stove, he or she could then see through its eyes and check if it
is on or not and take the appropriate action. All that would be
needed is an internet connection (and a robot, obviously).
A couple of things already stand out of this picture: It is a future
well-to-do single, autonomous (or isolated), emancipated individuals.
However, Brooks realizes that in such a system the human element
is (still) not out of the loop, and that in order for it to work
smoothly, and not disrupt the lives of those who rush to the airport
and have work to do during the day, someone will have to remotely
control the robot. The solution for this: developing countries where
labor is cheap (and often quite qualified). Remote work as the killer
application for robots in the short term. Brooks says, and I quote,
The brains of people in poorer countries will be hired to control
the physical-labor robots, the remote-presence robots, in richer
countries. The good thing about this is that the persons in that
poorer country will not be doing the dirty, tiring work themselves.
It will be relatively high-paying and desirable to work for many
places where the economy is poor. Furthermore, it will provide
work in those places with poor economies where no other work is
There are at least two ways of interpreting this statement. Brooks
may be a defender of the status quo, a pragmatist of sorts, and
have no faith that the power relations, that keep the system running
by forcing many entities to invisibility, can be changed. Brooks
could also be an idealist, which is certainly not a fault, and believe
that what he postulates above has a chance of becoming a reality.
In any case, he is remarkably uninformed about the working conditions
and lack of labour and safety regulations and working conditions
which many workers in developing countries are exposed to. One can
safely assume that he has not yet read Naomi Kleins No
Logo (2000), her exposé of the business practices of
large corporations and the realities of outsourcing and sweatshops
all over the world.
In the third, and last, section of the book Brooks turns to the
question of the essence of human and machines and the breaking of
the boundaries between both.
Brooks defends that the human body is a big, complex machine. The
difference between us and the robots, he says, is that we are [s]omewhat
more complex in quantity but not quality. In our overanthropomorphization
of humans, we tend to forget that we already have machines with
all the unique qualities of humans, we have ourselves. Thus, all
we need to do in order to create other life-like machine is to find
out what Brooks calls the juice of humanhood and replicate
it. To be sure, Brooks admits that presenting a mysterious juice
as the key to our existence is not very scientific. However, he
justifies himself by saying that contrary to the philosophers who
defend that this juice is something indefinable and uncopiable,
he is betting that the new stuff is something that is already
staring at us in the nose, and we just have not seen it yet.
The undertaking of the task of discovering this juice
is accompanied by efforts to promote the process of merging with
our machines, of creating man-machines. This is no novelty,
in our daily lives we have come to rely on technology to perform
many activities: ride the bus or drive the car to work, look for
information on the internet, write articles on the computer. The
social and the technical are equally central to our definition of
life, and we cross from one to the other without a second thought.
However, what Brooks is postulating here is that very soon these
extensions, to the moment primarily external, will be internalized
becoming part of our minds. We will have, for instance, direct mental
access to internet, or a thought controlled implanted mouse and
screen. The dream will then be complete, with humans becoming superhuman
in many aspects.
When this happens, Brooks predicts, [w]e will have the power
to manipulate our own bodies in the way we currently manipulate
the design of machines. We will have the keys to our own existence.
Flesh and Machines ends on this positive note of a superhuman
entity, constantly redefining itself and adapting itself to a new
world: A world where science and technology replace faith and provide
a new understanding of humanity.
I side with Brooks in his attempt to remove humans from their unique
status versus the-rest-of-the-world. In Western cultures, the bridging
of this gap is becoming increasingly visible due to the blurring
of the boundaries between humans and machines, for instance, through
the implant of artificial body organs. In other cultures the distinction
was never very strong, for instance, for the Achuar Indians nature
and society are part of continuum rather than independent spheres
Artificial intelligence, and by extension robotics, present themselves
as ideal locations to grapple with this rift. However, overthrowing
one authority to simply replace it with another is not a viable
solution. Asserting science as a privileged, universal, and thus
unifying, path to truth is not acceptable.
This brings us back to the beginning. To the initial criticism
of the oversimplified conceptions of technology (and science) presented
in this book. When technologies are reduced to tools, and science
to Progress, there is little space left for reflection or negotiation.
Science and technology become objectives in themselves, undisputed
However, there is a need to pause and reflect. In this text we
have explored at least two. The first pertains to the creation,
sustenance or expansion of social inequalities. The feeding
of the divide between those who have access to the technology and
those who dont, and the stifling of other sources of truth,
other forms of observation and interpretation. The second, relates
to the ways in which the creation of man-machines will
affect our experience of ourselves and the world. If, as Brooks
and Varela postulate, intelligence is indeed embodied and the world
is interpreted through our sensorimotor structures, then what are
the implications of adding, or removing, parts to this structure?
Brooks discusses these in terms of control, we are not in danger
of being controlled by our pure-robots, because we are like them
only better. However, there are other dimensions that deserve careful
reflection. If these are not considered, Brooks predictions
may become true for, as Marshall McLuhan once said, [w]e are
all robots when uncritically involved with our technologies
(McLuhan & Fiore 1968).
Unless otherwise noted all references are taken from Brooks (2002)
Brooks, Rodney A. (2002). <i>Flesh and Machines</i>:
How Robots Will Change Us. New York: Pantheon Books.
Descola, Philippe. (1994). The Society of Nature: A native ecology
in Amazonia (translated by Nora Scott). Great Britain: Cambridge
Latour, Bruno. (2002). War of the Worlds: What about Peace? (translated
by Charlotte Bigg). Chicago: Prickly Paradigm Press, LLC.
McLuhan, Marshall & Fiore, Quentin. (1968). War and Peace in
the global village. New York: Bantham.
Varela, Francisco J. (1999). Ethical Know-How: Action, Wisdom, and
Cognition. Writing Science Series. Stanford, CA: Stanford University