Would you pass the Turing test?

A lot has been written about Alan Turing this year; 23 June 2012 is the centenary of his birth. During his life, much of his work was secret and it’s only in the past few years that we’ve been able to appreciate fully the breadth of his contributions to mathematics, computer science, philosophy, and biology.

One of his achievements was to devise the Turing Test, a way of deciding if a computer could imitate human intelligence. In his test, a judge has a conversation with both a computer and a human, without knowing which is which. If the judge can’t distinguish between the computer and the human, then the computer is said to have passed the test.

This may have been the first time that the hard-edged objectivity of computers crashed up against our definition of humainty. The test relies on a rather subjective and individual judgment, perhaps different judges would arrive at different conclusions depending on their understanding and use of language.

And what happens to the human participant, how do they feel if they are accused of being a computer? Is a person’s sense of humanity sufficiently strong to not be undermined by a judgment passed on their behaviour in comparison to a bit of code?

Actually Turing first described the test as a sort of parlour game called the ‘Imitation Game’. This is carried out in two phases; in the first phase a man and a woman each try and convince the judge that they are female and the judge must choose who is right and who is wrong. Then a computer takes the place of the man and the test is repeated. So the test becomes a challenge to see whether a man or a computer can better imitate a woman. It’s bizarre to think that this could be a viable test of artificial intelligence, what does it mean to behave like a woman or imitate one in this sort of situation? Why should a man imitate a woman better than a computer? And given Turing’s own complicated (and ultimately tragic) personal life I can’t help reading into this theoretical test some sort of subtext about a wish to assume different sexual identities.

The test has been criticised on the basis that it turns on the imitation of behaviour rather than behaviour itself. I don’t know if this is a real concern, surely it’s possible for imitated behaviour to shade into learned or assimilated behaviour?

At one extreme you can have a ‘Chinese room’ type of imitated behaviour where the computer programme undergoing the test is merely following instructions without understanding them or having to adapt or change its responses according to the circumstances. But if the programme makes decisions in real time, or learns from previous circumstances, it’s hard not to see similarities with the way that humans learn behaviours.

I certainly feel like that about trying to write well, much of the impetus to write comes from reading other literature and trying to imitate what I like about it. In fact, when I was immersed in writing my novel, I had to stop myself from reading my favourite authors in case I started to mimic their traits in my own writing. There’s a fine line between imitation and forgery. What I read percolates slowly around my mind and through the unthought-thoughts.

And following on from my last blog post about Alan Garner, I was thrilled to find out that he thought of Alan Turing as his hero.

Posts created 145

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Related Posts

Begin typing your search term above and press enter to search. Press ESC to cancel.

Back To Top