Garry Kasparov is widely regarded not only as the best chess player now living, but perhaps the best who ever lived.
After agreeing to a rematch with the IBM computer called "Big Blue" that he defeated last year, he told a Reuters news reporter that he regarded the computer as an "alien opponent" he knows almost nothing about.
That "alien" defeated Kasparov on Sunday, May 11, 1997, ending a grueling six-game match watched around the world not only by chess aficionados but by droves of people who see in the computer's success the beginning of a new era for machine intelligence. Depending who you ask, this is cause for celebration, bewilderment or stark terror.
Kasparov told Reuters before the match began in New York that he had spent months training only with other chess-playing computers, trying to better understand and anticipate his silicon-based adversary. Going into the match, he exuded confidence. "I think that now my mind is working along the lines of computer 'psychology,'" he said.
"The computer is an alien opponent and the characteristics of this opponent are very, very different from any human opponent. In order to win the match, I have to look for moves that are most unpleaseant for the computer, moves that will take it off the right track," he said.
But even though he has successfully defended his world champion title eleven times against every top player in the game, he finally crumbled before the implacable Big Blue. Following the fifth game -- the third draw in a row -- Kasparov signaled his growing anxiety. "I'm not afraid to admit that I'm afraid...," he said. "It definitely goes beyond any chess computer in the world."
He resigned the sixth game after only 19 moves, getting up from the table visibly shaken and walking swiftly out of the room without comment.
Big Blue brought incredible computing muscle to the match -- an ability to review hundreds of millions of possible moves per second -- but that was not the edge that made it prevail. Big Blue was coached by another computer grand master named Joel Benjamin, who was hired as a full-time adviser to the five computer scientists working on the IBM project. Benjamin found ways to teach the computer how to think more like a chess master.
Chess is an activity ideally suited to a computer's strengths: the rules are rock solid, the variables limited, the goals clearly defined, the distractions minimal. Chess, in other words, is nothing like life on the street -- and Big Blue is a very long way from The Terminator. In fact, some experts say the only real impact of Big Blue's success may be to dampen the spirits of chess players.
Kasparov even alluded to the equivalent of beginner's luck when he said, after the match, "I feel confident that the machine hasn't proved anything yet. It's not yet ready, in my opinion, to win a big contest," by which he meant a tournament involving multiple opponents.
Others, though, are exultant. "One hundred years from now, people will say this day was the beginning of the Information Age," said C.J. Tan, head of the Deep Blue team. "Historically for mankind, this is like landing on the moon or being the first human to climb Mount Everest."
Still others, brooding on the astonishing advances in computing power over the last few decades, see a day not far off when a machine will literally "get smart." And that day, some believe, will not be a happy one.
This image, more often contemplated by science fiction writers than scientists, is epitomized by the superbly efficient killing machine brought to life by Arnold Schwarzenegger in "The Terminator." Many similar images come to mind: The Borg of Star Trek, for example, have wedded machine intelligence to a human frame, creating a hideously alien predator ("Resistance is futile"); the HAL 9000 of "2001, A Space Odyssey," is a highly personable machine that gets paranoid midway through a space mission and tries to kill everyone on board.
Of course, there are positive images as well. In "Terminator 2," Schwarzenegger's cyborg has been reprogrammed as a protector. Star Wars' C3PO, a "protocol droid," has a sort of aristocratic charm; his counterpart R2D2 is weirdly adorable. Author Isaac Asimov even imagined a brilliant android ascending to the presidency in his classic "I, Robot."
For better or worse, some theorists have concluded that the only extraterrestrials we are ever likely to meet are those suited to withstand the long travel times, inhospitable conditions and just plain boredom of interstellar travel -- in other words, intelligent machines.
But others see the advent of such machines much nearer at hand, here on earth. In a paper titled "Technological Singularity," mathematician and computer scientist Vernor Vinge predicts that "greater-than-human intelligence" will be created in 30 years or less. Vinge, a professor at San Diego State University who also happens to be an award-winning science fiction writer (his novel "A Fire Upon the Deep" won the Hugo Award in 1993), thinks that the "get smart" event might well surprise the humans who cause it to happen.
That moment, he says, will be "a singularity," and beyond that moment, the future of humankind is unknowable. "We will be in the Posthuman era," Vinge writes. "And for all my technological optimism, I think I'd be more comfortable if I were regarding these transcendental events from one thousand years' remove, instead of twenty."
Vinge is pessimistic about containing or preventing the advent of this "singularity." And once it happens, he says, all bets are off. "Just how bad could the Posthuman era be? Well,... pretty bad. The physical extinction of the human race is one possibility."
Needless to say, a chess-playing computer is no threat to the human race. But on this front, as on so many others, there are daunting ambiguities and unknowns. It is now possible to imagine that we humans will be the creators of the most formidible alien we could ever meet -- and soon.
Original file name: CNI - Deep Blue.Alien
This file was converted with TextToHTML - (c) Logic n.v.