The Turing Test

In 1950 the famous mathematician, logician, and cryptographer Alan Turing, proposed an altered version of an Imitation Game in his essay “Computing machinery and intelligence.” to be used for testing a machine’s capabilities in regards to intelligence.[1]

The Imitation Game is a game where two people, a male and a female, both pretend to be the female under interrogation from a third party.[2]  Turing envisioned both contestants sitting in separate rooms from the interrogator and that interrogation only take place through typewriting.[3]  The altered version of the game proposed as an Artificial Intelligence test would then take place with one of the contestants being machine and the other human, where a human interrogator would try to deduce which one would be the machine.[4]

My understanding of Turing’s argument paraphrased into an easily digestible version would be that ‘If a machine could be proven to exhibit mental features indistinguishable from those of a human being, we would have no logical objection to calling it intelligent.’ And his test is a proposed way of ascertaining such mental features.

There have been many critics of The Turing Test and its premises, with every possible objection imaginable ranging from ‘the test is too wide’ to ‘the test is too narrow’ or even ‘machines could never have souls.’ Some of these arguments Turing refuted himself while others, as for instance Searle’s Chinese Room - which we shall deal with shortly - are more recent and still being debated. Whether or not the test is valid, it has been adopted by many researchers worldwide as a set goal for AI to strive toward. This goal is particularly prevalent amongst developers of linguistic programmes, so-called chatbots, as for instance Eliza, Jabberwacky, and Alan Speak-A-Tron, just to name a few.[5]