Humans, take heart. Sure, Garry Kasparov, arguably the most gifted world champion of chess ever, lost in convincing fashion this month to I.B.M.'s supercomputer, Deep Blue (RS/6000 SP). But did the machine, with its talented satellite team of programmers and grandmaster consultants, cheat, as Kasparov suggested in his postgame comments? Common sense -- something that we should take more collective pride in than we do -- would tell us that of course the computer cheated, and, further, that no level playing field has ever existed between man and computer, at least not on the chessboard.
But Kasparov’s losing this match, and the wide-ranging philosophical and scientific implications that the loss holds for us -- implications that are at this stage probably more symbolic than substantive -- should compel us to examine more closely the unfair advantages (in Kasparov’s words) that we’ve been magnanimously allowing the machine to get away with. Perhaps we shouldn’t be so nice anymore.
Signaling this change of heart, Kasparov’s paranoiac charge that possibly a contingent of corporate-hired grandmasters had assisted Deep Blue behind a curtain during play was, nevertheless, the transparent reaction of a wounded ego. The accusation (“Look at the printouts!” he shouted) was eerily reminiscent of Bobby Fischer’s unrelenting paranoia about a Soviet conspiracy to keep him from becoming the world champion more than two decades ago (a paranoia that no one now dismisses as fantasy, in spite of the undeniable manifestations at that time of Fischer’s deepening psychosis). The better-adjusted Kasparov most likely will regret his comments when he cools off.
But some of his less caustic observations on the peculiarities of playing a supercomputer deserve to be put in a clear and more accessible light. For these more circumspect remarks of Kasparov’s, especially those alluding to the near-impossibility of gaining an advantage, much less holding one’s own, in the critical opening stages of a game with a computer, point precisely to just one of the absurd advantages that a supercomputer unfairly holds over a human opponent.
Perhaps the best way to illustrate this is to review the basics of fair play that most humans have religiously adhered to in the modern era of competitive chess tournaments (from the late 1800s, when the chess clock was introduced into serious competition, to the present day). In this context of what a level playing field means in chess, let us then compare how the pristinely amoral computer conducts itself during the course of a game.
The established and general rules of fair play in human chess competition are simple and commonsensical enough. Over more than a century’s time, they have helped to foster the misleading myth that it is impossible to cheat at chess, that the game is solely a mental mano a mano.
The rules demand that you may not consult with anyone (this would include a computer, these days) during a game; that you may not, in another version of the same sin, secretly refresh your leaky memory by consulting, say, a monograph on a given opening (e.g.,“the Nadjorf Sicilian”); and that you may not touch a piece until you are prepared to move it.
Further, each player is accorded equal blocks of"thinking-on-your-own-move” time. Traditionally, in tournament competition, this amounts to 40 moves per two hours for each player, as it did in Kasparov vs. Deep Blue. If players abide by the spirit and letter of these rules, indeed, a more level playing field between two opponents is hard to imagine. The better man, or woman, almost always prevails in the long run, hence the utter reliability of class rankings, from novice to master. These are the fundamental rules that have helped to elevate the game of chess to its elite status as"the royal game.”
But common sense tells us that chess computers routinely violate every tenet of this fair play. Whether Kasparov’s unseemly worries about grandmaster almost-peers-turned-corporate-lackeys illegally helping the team of programmers during play is true or not, the most obvious truth for all to see is that Deep Blue had at its disposal an"in-house” cyberspace library of chess literature that dwarfs virtually every known physical collection of books on the subject in three-dimensional space. Not only that, Deep Blue’s team of hired chess experts made sure that the library was kept up to date, and geared to respond to Kasparov’s pet openings.
Beyond this, it can be safely suggested that the supercomputer’s infallible memory banks make it possible for it to glean the results of more than a century of chess praxis in the well-charted waters of opening and end-game theory. This was why the disgusted Kasparov proclaimed that he was reduced to playing harmless"crap” in the opening. He had no reasonable alternative. As he well knew beforehand, it was in the infinitely more complex middle-game stage, after most of the pieces and pawns had been moved from their original squares, that his unique genius had the best chance of securing a winning advantage against the brute, calculating force of Deep Blue.
The computer reduces the other major tenets of fair play to absurdity as well. As long as computers play chess, nothing can be done about these unfair advantages, but they are worth mentioning anyway. With the computer’s capacity to examine millions of positions per second with breathtaking accuracy, it is, in effect,"touching” as many pieces as it likes throughout the game. Deep Blue has the ability to cheat more than 200 million times per second, a remarkable sleight of hand. So, too, it can cheat time by any mortal reckoning through the sheer speed and accuracy of its calculations. (The odds of a chess master’s beating Deep Blue at"speed” chess, which allows each player five minutes per move per game, are slim to none.)
What this all amounts to is that computers must cheat -- that is, break all the rules of fair play -- to humor humans with the illusion that the machine is"playing” a"game.” At least the computer must cheat if it is to play the game beyond the level that a 5-year-old could play.
For the time being, while philosophers, ethicists, and other specialists in the field of artificial intelligence debate the larger implications of Kasparov’s loss, perhaps we should, like Kasparov, indulge in a little petulance ourselves.
Glenn D. Klopfenstein is an assistant professor of English at Passaic County Community College. He has been designated a “life master” and"expert” player by the U.S. Chess Federation.