Stay up to date with notifications from The Independent

Notifications can be managed in browser preferences.

Chess

William Hartston
Saturday 29 November 1997 00:02 GMT
Comments

How can you tell a human being from a computer? When Alan Turing proposed his famous "Turing Test" for machine intelligence his thoughts were clearly on too civilised a plane to encompass the idea of people cheating in a chess simultaneous display.

His idea was to have an experimenter asking questions of two respondents, one human, one machine. If the computer can be programmed so that its responses cannot be distinguished from the human's, then, by Turing's criteria, one has to admit that the machine is intelligent.

So what if the experimenter is the world chess champion, and his questions take the form of moves played in a series of chess games? It is not difficult to devise a set of positions that will discriminate between computers and humans, but if the champion's primary objective is to win the games, then discrimination becomes more difficult.

So what was it that made Garry Kasparov think that a team competing against him over the Internet in a simultaneous display a couple of weeks ago was using a computer to help select its moves? Black's excellent opening play - and the presence of a computer beside the board (necessary for the Internet connection) - may have been enough to alert Kasparov's suspicions. The calmness with which they then threaded their way through complex tactics seemed to confirm it. Finally, 29...Rf8! was too much for him. After 30.Rxf8+ Kxf8 White has no good discovered check with his knight, because the threat of Qc1+ is hanging over him.

So Kasparov refused to continue the game. The captain of the Internet team has strongly denied all charges, pointing out that the players were being watched all the time. I don't know who is right. See what you think:

White: Garry Kasparov

Black: Internet Chess Club

1 d4 Nf6 16 Ng5 e3

2 c4 g6 17 cxd6 cxd6

3 Nc3 Bg7 18 fxe3 Bh6

4 e4 d6 19 Rxf8+ Qxf8

5 Be2 0-0 20 Ne4 Bxe3+

6 Nf3 e5 21 Kh1 Nf5

7 0-0 Nc6 22 Ra3 Bd4

8 d5 Ne7 23 Qd2 Qd8

9 b4 a5 24 Bg4 Ne7

10 Ba3 axb4 25 Bxc8 Qxc8

11 Bxb4 Nd7 26 Nxd6 Qd7

12 a4 f5 27 Rf3 Nxd5

13 a5 Nf6 28 Qa2 Qc6

14 c5 Nxe4 29 Rf1 Rf8

15 Nxe4 fxe4

Join our commenting forum

Join thought-provoking conversations, follow other Independent readers and see their replies

Comments

Thank you for registering

Please refresh the page or navigate to another page on the site to be automatically logged inPlease refresh your browser to be logged in