Key Flaw of the Turing Test

The most obvious and crippling flaw of the Turing Test is that Turing has no way of proving that machines can think, any more so that his opponents can prove that they cannot think. However, this percieved flaw is due to misinterpretation of the Turing Test's scope. Turing's original argument was not that the Turing Test be used to "prove" that a machine is human, rather as a test of a computer's capabilities. His statement, "We may hope that machines will eventually compete with men in all purely intellectual fields," (Turing) sums up his intent. In that light, Turing's response to the Argument from Consciousness makes sense: if the test is a proof of ability, if the test subject demonstrates that ability, then it passes the test. At least in his original paper, Turing is not attempting to equate machines with humans on a metaphysical level; such speculation is outside the scope of his paper.

Where Turing gets into hot water is his usage of the words "intelligent", "think", and "consciousness" to describe a Turing-capable computer's abilities. Perhaps because of his own beliefs as to the eventual destination of computer development, Turing is quick to attach these words to machines, even as he urges that machines and humans are fundamentally different and therefore should not be compared in all respects. His intent with the test was perhaps to use these terms to convey the unknowable nature of intelligence; several of the objections he addresses simply reject the idea of a thinking computer as blasphemous or unthinkable. His test questions that idea by presenting a case in which a computer cannot be distinguished from a human; he asks, if that is not intelligence, what is?

The Turing Test asks hard questions about the nature of intelligence, and must use terms that lead to controversy. Many today have the mistaken assumption that the Turing Test is intended to prove that a computer is human, an assumption that ignores the scope of Turing's conclusions.