Robotics Reference
In-Depth Information
AI has been thought controversial because it challenged the unique-
ness of human thought, as Darwin challenged the uniqueness of human
origins. The boundaries of AI continue to expand rapidly, settling the
controversy for those who know the evidence. [3]
The Turing Test
The Imitation Game described in Turing's paper has since become the
basis for what is known as the
Turing Test
. This phrase has been used
more generally to refer to some kinds of behavioural tests for the pres-
ence of mind, or thought, or intelligence in entities that are intended to
appear intelligent. The modern form of the Turing Test, as employed for
example in the Loebner Prize competitions, calls for human judges to be
given the opportunity to interact via a computer terminal with whatever
entity is at the other end of the communications link, be it human or
computer program. Typically each judge is given a ten-minute session
on each terminal and, once all of the terminals have been interrogated
by all the judges, each judge is given a further five minutes with each
terminal. The judges may interact in any way they wish, asking whatever
questions and making whatever statements they like. At the end of the
day every judge ranks the terminals from “most humanlike” to “least hu-
manlike” and the program with the highest average ranking is the winner
of the competition. Each judge is also asked to draw a line in his or her
ranking list—above this line are those terminals which the judge believes
to have a human at the other end.
This form of the test is very much as Turing himself suggested, using
conversation in natural language as the domain in which to determine
whether a computer program can justifiably be said to be intelligent. If
a program can convince the judges that it is “above the line” (i.e., hu-
man), then that program will pass the Turing Test. But it is not only
in the realm of human-computer conversation that the Turing Test ap-
plies. Indeed, it is relevant to just about every realm in which human
intelligence is a prerequisite for performance at a high level. An example
from a different realm is Chess. As the twentieth century drew to a close
the world's leading Chess programs performed at the upper echelons of
human playing strength. It seems perfectly reasonable to argue that, by
achieving, in this highly intellectual pursuit, the level of the world's top
human players, the best programs had passed the Turing Test for Chess
because their standard of play was so high. Yet Kasparov and a few other
