The Turing Test: The Elusive Standard of Artificial Intelligence

Alan Turing begins his 1950 Mind article, “Computing Machinery and Intelligence,” with the following straightforward pronouncement: “I propose to consider the question ‘Can machines think?’ ” He quickly (too quickly?) argues for “replacing” the original question with another “which is closely related to it and is expressed in relatively unambiguous words.” (Please note the vagueness of the claim; what is not vague is that it is not a claim of identity.) Turing goes on to describe the new form of the “problem” in terms of an “imitation game,” played with three people, a man A, a woman B, and an interrogator C, who may be of either sex. They are placed in separate rooms, with the only form of communication between them being teleprinter (or, to be au courant, instantmessenger clients on their desktops). C is to ask them questions or otherwise engage them in conversation, with the object of the game being for C, who knows them only by the labels X and Y, to determine which of them is a man and which a woman. Turing then proceeds to offer his replacement(s) for the original question: