h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

The Turing test is a test used to test how good an AI system is at immitating humans. It works something like this :
If you were to ask a set of questions to a computer screen by typing them in could you determine from the answers who was writing it ?
A computer could pass the test if you could not say if the repsonse's come from a human or not.
Hope this makes things clear
(knew that degree thing would come in handy somewhere!)

It was named after Alan Turing, a mathematician who during WW2 came up with a machine to decode the German coded messages. The machines were the Bombes, and Collosus, primative computers.
After the war he cantinued research into computing and AI and he devised the Turing Test, which Scarlett explained before.

Thanks for the explanation

There are those (and I tend to agree with them) who would say that the Turing test isn't a measure of whether AI has 'made it', as it only tests a machine's imitation of one aspect of human behaviour.

Imagine a wall with a hole at ground level. A shoe sticks out of the hole. You drop a box on the shoe, and the shoe disappears into the hole. The shoe reappears, you drop another box on it and it disappears into the hole again. The shoe reappears, you drop a third box on it and this time another hole opens in the wall, and a fist punches you in the mouth.

Now that might have been a person getting irritated about having boxes dropped on their feet, or it might have been a computer that was programmed to respond in that way; you can't tell. If it was a computer, then it has successfully mimicked an aspect of human behaviour. The question is whether that aspect, despite being easier to model, is as valid as conversation. I tend to think it is.