Science

25 years ago, the computer tamed the chess king

Right hand under his chin, dark eyes, Garry Kasparov looks at the chessboard one last time before abruptly leaving the table. This is a thunderclap: the chess king has just been beaten by the computer. On May 11, 1997, for the first time since the main match, the car knocked down the reigning world champion. This date will go down in the history of the discipline and shed light on the dizzying potential of artificial intelligence.

Champion lost after six games to IBM’s Deep Blue

Baku Ogre, 34, and then master of the world chessboard since 1985, lost after six games with a final score of 3.5 points to 2.5 against IBM’s Deep Blue supercomputer. In front of cameras from all over the world who came to shoot the show in New York, this failure smacks of humiliation for the exuberant Kasparov. Didn’t he promise that he would stand against the machines until at least the dawn of the next millennium?

A year earlier, the Russian beat Deep Blue with a score of 4:2. But its designers constantly improved the 1.4-ton monster, which is now capable of counting 200 million positions per second.

“The computer hasn’t proven anything yet”

Shocked by his defeat, the champion nevertheless refuses to acknowledge the superiority of the machine. “The computer hasn’t proven anything yet,” he said at a post-match press conference, according to AFP. “The human being, the best player in the world, broke down under pressure,” he explains, referring to himself, “but the computer can be beaten, it has too many weak points.” Far from taking comfort in the $400,000 promised to the loser, he criticizes IBM for not giving him access to previous games played on a machine that was able to analyze all of its own.

He even blames himself, suggesting that people helped the computer during the match and regretting that he did not “set certain conditions” for the game to be “fair”.

The rest of the world’s top chess players, who were closely watching the confrontation live, also refuse to consider the grandmaster’s defeat a turning point. When asked by the press, they point to a number of unsuccessful decisions by the Russian champion. For some, his obsessive need to understand the punches of the machine rather than focus on winning was fatal to him.

Years later, the book would reveal a secret from developer Deep Blue: a computer glitch could have crashed the game. Without being able to choose between several moves, the machine would play randomly during the game, destabilizing Kasparov for the rest of the confrontation.

“A little in hindsight,” his defeat seemed to him “a victory for the human race.”

In any case, Deep Blue’s victory pleased IBM, seeing interest in its computing capabilities. “This has nothing to do with man vs. machine, but how we as humans can use technology to solve complex problems,” said IBM Project Manager Chang-Jen Tan enthusiastically after the match. And to praise what artificial intelligence will bring to many areas, from financial analysis to the study of natural, meteorological or seismic phenomena.

Kasparov, revenge, in 2003 will draw twice with computers. But time will eventually soothe his wounded ego. After the 1997 defeat, “I was devastated,” he admitted in a 2019 interview with the Swiss newspaper Le Temps. “But in retrospect, his defeat seemed to him a ‘victory for humanity’ as it anticipated ‘the breadth of the spectrum of activities that technology can help’.

From now on, the chess legend worries about digital giants overstepping personal freedoms: “We need them to be held accountable,” he urged in an interview given to AFP in November 2021.

Back to top button