Information Technology Reference
In-Depth Information
even theoretically possible? Details of human thought remain a mys
tery, so much of the answer to this question depends upon your be
liefs regarding the nature of human intelligence. If you believe that
humans behave largely based on learned responses to events and
stimuli, then you might well believe that research will slowly, but
methodically, unpeel the workings of the human brain. Eventually
this could lead to understandings of how we think, and such knowl
edge could lead to various mathematical models and computer ap
plications. From this perspective, computers might eventually be able
to think, because we will have learned enough about human thought
to model or simulate it.
On the other hand, you may believe that human thought in
volves something more than electrochemical activities. Until this
something more is identified, it would seem that computers might
not be able to simulate it for whitebox testing of intelligence—al
though there still might be sufficient technology to handle much
blackbox intelligence testing. Although this perspective is appeal
ing to many people, it seems to be under constant attack as we learn
more about human reactions (as in the experiments of Kornhuber),
and as we try to identify what other element(s) might be missing.
Certainly, in the coming years, we all will learn more about the
nature of human thought, and many new technologies will emerge
to solve problems that we have associated with intelligent behaviors
in the past. It will be exciting to observe whether these areas of hu
man thought and computer applications converge, or whether they
continue to follow largely separate tracks.
Consideration of whether computers might one day think begins
by investigating what we mean by intelligence. Although some simple
definitions can trivially include or exclude the notion that computers
could think, these definitions are not very helpful. For example, in a
careful analysis of the nature of intelligence, it is not clear that either
consciousness or free will is a necessary characteristic.
To avoid biased definitions, objective assessments of intelligence
could use either blackbox or whitebox testing. The Turing Test is
the most famous of the blackbox testing methods for determining
whether or not computers might be able to think. Some criticisms
