This is an abstract of Alan M. Turing’s “Computing Machinery and Intelligence” paper, which outlines what is, in many circles, the accepted standard test to determine if a machine is intelligent. You can read the original paper on the Hugh Loebner’s site, here. It’s a straightforward read, even if you aren’t a computer scientist.
I have my doubts about the test–it winds up being more a test of emotional intelligence than anything else–but this is a must-read if you want to understand what has motivated AI people in the past. The test continues to be relevant to many; Ray Kurzweil has placed a wager that a machine will be able to pass the test by 2029.
The question “Can machines think?” may most easily be decided by seeing whether they may imitate a human so well that we fail to tell them apart, says Turing. To play this “imitation game” fairly, we must remove physical factors so far as possible, and we should consider only digital computers (no “biological machines” like human clones allowed). Since we may implement an exceedingly large number of programs on a computer, this is but a minor limitation. While there may be some machine able to think but unable to successfully play the imitation game, we may say with certainty that any machine able to play successfully does think. Various objections, based on mathematics (“Logical systems of complexity are necessarily incomplete”), consciousness (“Computers can never feel“), learning (“Computers can only do what you tell them to”), and more are considered, but all are ultimately rejected as being insufficient to refute the test.