Welcome to the A.L.I.C.E. AI Foundation

Promoting the adoption and development of Alicebot and AIML free software.

Turing's Machine

Gregory Khasin

Gregory Khasin

Wondering whether a machine can think, the English mathematician Alan Turing described the following thought experiment. He imagined a locked room with a computer inside. Questions can be fed into the room, and its hidden inhabitant must reply. If, based on such a dialogue, we cannot determine whether the inhabitant is human or machine, then the machine can think.

Let's reflect on this definition. Turing's idea is simple: if a machine can deceive us, it can think. For machines there is no difference between thought and its simulation.

Let's take the next step. Turing is dealing not with a technical problem, but with thinking in general. He's asking whether machines can do what people do. It follows that the concept of thought which he applies to machines must be applicable to humans as well--otherwise his definition is useless. According to this line of reasoning, if simulated thought is equivalent to real thought for machines, it must be also so for us. We think, Turing implies, in virtue of our ability to fool each other successfully.

Thinking, then, is cunning. Unconscious, of course--the sober Englishman would never accuse reason of premeditated treachery--but ubiquitous and fundamental. We tolerate each other and work together only thanks to a complex system of mutual deceit, in which everyone plays smart for everyone else.

Such a theory makes us look at the world with suspicion. It's painful and scary to find oneself in a sea of lies. History records many attempts to uncover how things really are, to break into that damned room. Alas, that's impossible. The consequence of Turing's definition is that thinking takes place when and only when we cannot check whether it is taking place or not.

As a result, we have to rely on the honesty of our fellow men, or, rather, on their dishonesty, on their ability to give the right answers. Many break down, tired out by the demands of this constant make-believe, and we have to force them back into the bosom of society. A person, to be considered human, must be able and willing to deceive his neighbor. If he's not up to the task, the neighbor must help him--give advice, look the other way, keep up appearances. Even if someone, driven by idiocy or turpitude, betrays his true nature (which, frankly speaking, happens more and more often nowadays), we must cover it up. Under no circumstances can we let him guess that we know more about him than he knows about himself. The art of fooling others into believing that they're fooling us is the basis of civilization. It's terrifying to think what would happen to us had family and school not initiated us into this art at a tender age. A true member of society, a citizen and gentleman, must pretend to be human so well that even he himself cannot see his own bluff. Then no machine can come close to him, since machines, dumb things that they are, cannot deceive themselves, however hard clever Turings may try to help them.

Gregory Khasin
Universtity of Chicago, Slavic Studies, Ph.D.
St. Petersburg Polytechnical Institute, Thermal Physics, M.S.
email: gkhasin@yahoo.com