LEADING ARTICLE: Between a rook and a hard place

Click to follow
The Independent Online
A small victory for a machine, but a giant leap for silicon-kind. When Garry Kasparov, the world's most formidable chess player, resigned the first game of his match against the IBM computer, Deep Blue, on Saturday night, the world of artificial intelligence celebrated. Ever since Claude Shannon in 1950 outlined how a chess computer program could be constructed, artificial intelligence has sought to build a machine that could defeat the world's strongest grandmasters.

Early predictions of the imminent downfall of the human chess brain, however, proved to be absurdly optimistic. Now the frequent assertions of eternal supremacy from the other side of the chessboard have been shown to be equally mistaken. Kasparov may yet come back from this defeat to win the match against Deep Blue, but few can doubt that the days of human chess supremacy are numbered.

A machine beating a man at chess is not important in itself, but raises the vital question: if chess, what next? What other problems offer a domain of similar size that we might now expect computers to turn their processor towards solving? Two years ago, Garry Kasparov saw his role as a champion of human intelligence. He said that if computers could play chess better than people, then they would be able to write novels and compose great music. Yet the domains of language and music are far greater than that of chess. Computers may already be able to write limericks, but a machine- written War and Peace or an Eroica symphony are a long way off.

That said, the fields of decision-making such as medical diagnosis, air- traffic control and economic forecasting may all be of roughly the same order of magnitude as the game of chess. Perhaps the most important aspect of Deep Blue's achievement will be its role in educating us to respect the judgement of machines in certain areas that have traditionally been entrusted to humans.

The result may also, perversely, increase our respect for the human mind. With up to 32 pieces scattered over 64 squares, a chess position has a level of complexity that ought, by most criteria, to be beyond human calculation. It is encouraging that it has taken a monster of 256 processors, working out a billion moves every second, to beat it. Perhaps we humans are not so dumb after all.