I can’t remember the genesis of this particular article on the Turing Test, but this year being the centenary of Alan Turing’s birth certainly makes it very apt a year later.
In essence, in the early 50s Turing was considering the question: can machines ever be said to think? He was at the time involved in the building of the first general-purpose computer and, based on his work with the theoretical programmable machines now known as Turing machines, was pondering how intelligent these computers could get. Instead of answering the original question, he turned it around: instead of asking can machines think, he wondered, given a communications device, can a human (who, for all intents and purposes, can be assumed to be a thinker) detect whether the counterparty at the other end of the communications link is human or machine?
From this interesting premise, we now have much faster computers and naturally people have programmed them to try and pass the Turing Test. ELIZA was the first of these programs, but now we have “chatterbots” on various websites to provide a initial point of contact for support. (PayPal uses a program called Sarah, for example.) In turn these chatterbots have led to much improved natural language parsing techniques, which then can “understand” natural conversations, which in turn make the chatterbots more life-like.
All in all, we’re not quite there yet as regards a program that passes the Turing Test, but we’re getting better at programming it and closer to a thinking machine. At least in the Turing sense.
This article first appeared in issue 313, October 2011.
You can read the PDF here.
(I used to write a monthly column for PCPlus, a computer news-views-n-reviews magazine in the UK, which has since been closed down. The column was called Theory Workshop and appeared in the Make It section of the magazine. When I signed up, my editor and the magazine were gracious enough to allow me to reprint the articles here after say a year or so.)