Paul Vallely: The perpetual struggle of man against machine

Kramnik was not playing a machine, he was taking on the ghosts of grandmasters past

Don't talk to me about man versus machine. It was a grievous disappointment, but hardly a surprise, to hear over the weekend that the world chess champion, Vladimir Kramnik, had failed to salvage humanity's honour by beating the world's top chess computer, which exulted in the name of Deep Fritz. He managed only a draw, but that was hardly compensation for the previous contest between brain cells and silicon when the former world champ Garry Kasparov was whupped by a super-computer in 1997.

When it comes to mechanised assaults on human dignity, I know a thing or two. I tried to book some train tickets the other day. After 25 minutes on the internet pitted against some programme incapable of understanding why anyone would want to buy a ticket for a child under five (because you have to if you want adult discounts on a family railcard), it eventually became clear that the programme at the other end was no Fritz, even a shallow one. So I gave up on-line, and telephoned.

To no avail. A machine answered that too, and a voice-recognition programme started to ask me questions and then, when I replied, told me: "I can't hear you." Fourth time round I was reduced to bawling with undisguised ire: "I want to buy a bloody ticket." I slammed down the phone before some Virgin Rail anti-obscenity filter could cut in.

The trouble with artificial intelligence is that it is not intelligent at all, but something else masquerading as cleverness. Deep Fritz apparently is capable of evaluating six million moves a second. Genius chess grandmasters, by contrast, can manage just two or three. World champion Kramnik modestly suggests that he can only manage one a second, but then adds "but always the best one." To write down the number of unique games possible in chess you would, evidently, have to write 120 noughts after the number 10. Given that, you would have thought that the brute force of a super-computer's calculating ability would make it unassailable. And yet there is evidently something about the quality of Kramnik's intuitive response which is at least equal to the might of the machine.

The truth is, of course, that in reality computers do not emulate human methods of thinking. How could they, since we don't really know ourselves how the brain works, with its unfathomable complex labyrinth of interlocked neurons with processing and memory distributed throughout? Machines can perform some specific functions better, faster and more accurately than we can. But no formula exists for intuition, let alone wisdom.

What Deep Fritz combines is this relentless capacity for running through options together with the programme's compilation of the wisdom of the game across the centuries. In one sense, Kramnik was not playing a machine, he was taking on the ghosts of grandmasters past.

But there is another key difference. Kramnik and Deep Fritz played eight games. The first was a feet-finding draw. Then the human easily won the next two, never appearing in any danger as he exploited the playing style of the computer, which is essentially responsive and lacks any real concept of long-term planning as well as any intuition. Kramnik's stratagem was to frustrate its ability to look ahead and predict which way the match was going. But then he made an elementary blunder in a difficult position to lose game five. And, perhaps unnerved by that, he then resigned in game six, in a position that other grandmasters thought was still tenable. His response was to play safe in games seven and eight, going for short draws.

A different kind of a calculation was at work here. He knew that going for a win also meant risking a loss, so he played safe. Chess players, like all human beings, know when they are not at their best, and Kramnik had clearly decided that if he was not going to get the winner's $1m purse he would rather have the $800,000 for a drawn match rather than the loser's consolation of a mere $600,000. "Chess is 30 to 40 per cent psychology," the game's leading woman player, Judit Polgar, said after losing to a software package called Deep Thought II. "You don't have this when you play a computer. I can't confuse it."

Chess, some sage once opined, offers a mirror to self-understanding. It is not just about intellect but about handling a mixture of pleasure, elation, fear and even hatred. ("I hate anyone who beats me," one grandmaster memorably said.) A computer has none of that. It can neither intimidate nor be intimidated. It experiences no joy when it wins or sadness when it loses. The most galling thing about the Virgin ticket line was not the impenetrable stupidity of the machine. It was the sound of my wife chortling in the background. Even when she fled to the bathroom to disguise her mirth, I could still hear the muted sound of hysterical laughter. Bloody machines.

p.vallely@independent.co.uk