Deep Blue Beats Chess Champion

In May 1997, an IBM supercomputer called Deep Blue defeated world chess champion Garry Kasparov in a six-game match in New York City. It was the first time a reigning world champion lost a match to a computer under standard tournament conditions. The result became a cultural landmark, widely taken as a sign that machines could rival human experts in a task long associated with intellect, though it also sparked debate about what such a victory really showed.

Computer chess had been a goal of researchers since the dawn of computing. Pioneers such as Claude Shannon and Alan Turing sketched chess-playing methods in the 1940s and 1950s, and chess became a favorite test for artificial intelligence. Deep Blue grew out of work begun by graduate students at Carnegie Mellon University, who were later hired by IBM. Kasparov had already played the machine's predecessor, Deep Thought, and in 1996 he faced an early version of Deep Blue in Philadelphia. He won that match, though the computer took one game, the first time a machine beat a reigning champion in a single classical game.

For the 1997 rematch, IBM's team substantially upgraded the system. Deep Blue combined custom chips built to evaluate chess positions with a large parallel computer, allowing it to examine an enormous number of positions each second. Its evaluation was tuned with the help of grandmasters, and it drew on stored opening and endgame knowledge. Kasparov won the first game, lost the second, and then drew the next three. Deep Blue took the sixth and final game, giving it a narrow overall victory. Kasparov was visibly shaken, particularly by the second game, in which the computer's play struck him as surprisingly subtle. He suspected human intervention, an accusation IBM denied. Debate over the match's conditions continued, and IBM declined a rematch and retired the machine.

Deep Blue's win did not mean that computers had become generally intelligent. It relied largely on brute-force search of possible moves, not the flexible reasoning or learning that characterizes human thought. Many researchers noted that it was a specialized tool, highly effective at one narrow task. Still, the match marked a turning point in public perception of what machines could do.

Chess engines have since become far stronger and are now standard training tools for players at every level. Later milestones, such as AlphaGo's 2016 victory over a top Go player, used machine learning rather than mainly exhaustive search. Kasparov himself later advocated pairing human players with computers, and the 1997 match remains a defining moment in the history of artificial intelligence.