Robotics Reference
In-Depth Information
We now ask the question, “What will happen when a machine takes
the part of A in this game?” Will the interrogator decide wrongly as
often when the game is played like this as he does when the game is
played between a man and a woman? These questions replace our
original, “Can machines think?” [2]
Later in the paper Turing discusses his own belief, namely that the answer
tohisquestionis“yes”.
It will simplify matters for the reader if I explain first my own be-
liefs in the matter. Consider first the more accurate form of the
question. I believe that in about fifty years time
6
it will be possible
to program computers with a storage capacity of about 10
9
to make
them play the imitation game so well that an average interrogator
will not have more than a 70 per cent chance of making the right
identification after five minutes of questioning. [2]
Turing's figure of 10
9
is 1 gigabyte in today's terminology, roughly the
memory capacity of a typical PC's hard disk in the closing years of the
twentieth century. By the year 2000, although computers had not been
programmed to enable them to play the “imitation game” so well that
an average interrogator had no more than a seventy percent chance of
making the correct identification after five minutes of questioning, some
of the programs entered into the annual Loebner Prize competition
7
are
not so very far away from this goal. My own program Converse, which
won the 1997 Loebner competition in New York, convinced one of the
five judges after his first ten-minute session with the program that it was
in fact human, only giving itself away in a subsequent five-minute session
later that day.
There are other predictions made in Turing's 1950 paper that already
appear to have been vindicated. In particular, educated opinion has al-
ready been altered to the extent that one can now speak of the possibility
of machines thinking and learning, and the term “machine intelligence”
is not the oxymoron that it might have been taken for when Turing first
started thinking about the subject.
Turing's paper continued:
The original question, “Can machines think?” I believe to be too
meaningless to deserve discussion. Nevertheless I believe that at
6
I.e., around the year 2000.
7
See the section “Passing the Turing Test” in Chapter 7.
