Turing’s Test

Seven years after McCulloch and Pitts suggested that the brain was a digital computer, Turing tackled the question of whether machines can think in his 1950 article “Computing Machinery and Intelligence” (Turing 1950). Suppose we play an imitation game with a computer and two people. One person—the interrogator—types questions and sends them to the participants. Seeing neither participant and just receiving text answers, can the interrogator identify the computer? If, on the one hand, the interrogator easily does so, we have failed to create intelligence, as the computer’s behavior fools nobody. If, on the other hand, the interrogator frequently guesses incorrectly, we should infer that the computer is intelligent. But why should we infer that the computer is intelligent merely because it behaves as if it was?

Turing reasons that we do this with humans—we behaviorally infer their intelligence. One neither experiences the minds of others nor knows much about their brains, but this doesn’t hinder ascription of intelligence. We shouldn’t treat computers differently. Turing discussed AI with Shannon when he visited Bell Labs and provocatively exclaimed:

I’m not interested in developing a powerful brain. All I’m after is a mediocre brain, something like the President of the American Telephone & Telegraph Company. (Hodges 1992, p. 251)

Let us look further at some of Turing’s work to understand better how he viewed the possibilities of intelligent computers and how AI and cognitive science evolved.

< Prev   CONTENTS   Source   Next >