A Critique of the Turing Test

Home ] Up ]


Critics of the Turing Test, which is concerned with the imitation of human thought processes with computer algorithms, do not need any assistance from me. However, there are some simple observations which cast some doubt on the whole subject, and which I would like to present. Bear in mind that the specific form of the Turing Test has changed many times over the years and has been expressed in different ways. Nevertheless, the central idea is to create an environment in which observers must judge whether the interactions they witness are from a human or a computer. If a large number of observers are 'fooled' by a computer program, the program is said to imitate thinking.

Wizard of Oz

Consider if this same type of argument were posed to citizens of Oz. They, presumably, believed they were in the presence of a Wizard. It wasn't until someone (Toto) pulled back the curtain that it was revealed that the Wizard was a hoax. Isn't this essentially the same test that Turing proposes?  Of course, Oz is a fictional city. Still...

A New Analogy

As a new analogy to the Turing Test, I propose the Bond Test. This test attempts to determine whether an object is alive or not. It consists of the following steps:

  1. Construct a mechanical mouse complete with artificial fur, whiskers, beady eyes and a tail,
  2. Wind it up and release it in a room full of secretaries,
  3. If any of the secretaries jump on their desks, the mouse is alive.