Wednesday, August 19, 2015

[ESSAY] The Kasparov Effect

"I'm sorry Frank, I think you missed it. Queen to bishop three, bishop takes queen, knight takes bishop, mate."- HAL 9000

February 10, 1996. The first time that a computer defeated a reigning world champion of chess (in this case, Garry Kasparov) using standard tournament rules. Deep Blue won Game 1 of the match. Speaking of his opponent, Kasparov stated "My late game attack would intimidate many players into making a mistake or two, but not this one." His coach would go on to tell TIME magazine that the defeat was "a shattering experience" for Kasparov. Garry Kasparov would go on to win 3 games and draw 2 of the remaining games in the match, ultimately concluding 4-2, Kasparov-Deep Blue.

Kasparov and IBM would go on to schedule a rematch for 1997. That match would ultimately end in Kasparov's defeat and the first match defeat of a reigning world champion by a computer using standard tournament rules.

Deep Blue versus Kasparov, Game 1 ended in victory for Kasparov. But Kasparov was mystified by Deep Blue's 44th move and subsequent resignation. He attributed Deep Blue's move to "superior intelligence" when, in all likelihood, it was a fail safe move (or a bug, according to WIRED). The program, when confronted with no optimal solution, merely chose one at random from all possible moves. Garry Kasparov's belief in the "superior intelligence" of Deep Blue would go on to hurt his game in subsequent matches.

Deep Blue versus Kasparov, Game 2 ended in victory by resignation for Deep Blue. Kasparov, showing signs of distress, sighing and rubbing his face, walked away and resigned after the 45th move, only to later be told that he could have fought the endgame to a perpetual check (and therefore, a draw.) "If he had not stormed off the stage and just played his normal game, he could've tied Deep Blue" (NPR) Garry Kasparov had let himself be psyched out, by a computer he would later say was "only intelligent the way your programmable alarm clock is intelligent." 

Games 3, 4 and 5 ended with mutually agreed draw. Going into Game 6, the match was tied 2 1/2-2 1/2. There was yet hope for Luddites everywhere. It wouldn't last.

Deep Blue utterly defeated Garry Kasparov in that game, concluding in Kasparov's resignation by the 19th move. The whole game lasted under an hour. Naturally, Garry Kasparov, being a good sport and a gentlemanly competitor, accused the IBM Deep Blue team of cheating. He had come to believe that Deep Blue was being guided by a human player, a la the Mechanical Turk, despite his earlier recognition in the first match between himself and Deep Blue that the computer could play "wonderful and extremely human moves". At the post-match press conferences, he was described as "puffing and pouting". Kasparov would continue to accuse the IBM Deep Blue team of cheating before eventually just refusing to discuss his defeat against Deep Blue altogether.

Why did Garry Kasparov let himself get so flustered up against a machine opponent? Why was he such a sore loser after the fact? And what relevance does this have to the wider world?

Deep Blue was not necessarily a better chess player than Garry Kasparov. But unlike Garry Kasparov, Deep Blue was just a chess player. Garry Kasparov was a human with emotions and the possibility to err. More than Deep Blue defeating Kasparov, Deep Blue enabled Kasparov to defeat himself. As Mike Greengard, Kasparov's publicist and confidant, said "Deep Blue could calculate 200 million possible moves per second, but it was Kasparov who is overthinking it."

Garry Kasparov was not the first and will not be the last to be profoundly demoralized upon contact with machine intelligence. Inside Chess magazine's cover following the Kasparov defeat read ARMAGEDDON! 

As the ubiquity of machine intelligence in the world only increases, we can expect only more anger and dejection, more feelings of defeat and coming doom. The wide range of human endeavors that machines are now invading will provoke a very raw, human response- a Kasparov Effect, if you will.

What will be the response to the first human casualties due to self-driving cars? Will the automation of jobs provoke more public rage and sympathy than the outsourcing that gutted American manufacturing and others? Should anger at the defeat of man by machine be a legitimate concern in public policy? Is the out-of-proportion emotional response these things will provoke justifiable and worthy of consideration? These are all questions that must be answered and stem from the same bewilderment that Kasparov faced in his rematch with Deep Blue in 1997.

Perhaps it is instructive to look at Garry Kasparov's statements before his second match against Deep Blue. He said "Inevitably the machines must win, but there is still a long way to go before a human on his or her best day is unable to defeat the best computer."

Inevitably the machines must win, indeed. How long a long way is remains to be determined.

No comments:

Post a Comment