I watched Wednesday night's episode of Jeopardy! with someone who's a three-day winner and Tournament of Champions player -- not to mention a staffer here at THE WEEKLY STANDARD. A few of us switched it on in the office not to watch Jeopardy! legend Ken Jennings, though he was a player that night, but to see Ken Jennings and another contestant, Brad Rutter, beat by a machine. "I, for one, welcome our new computer overlords," the clever Jennings wrote underneath his (correct) answer in the final round. But should the rest of us?
I do wonder if we should be so certain we will one day be ruled by machines. Watson, a "supercomputer" developed by researchers at IBM and named after company founder Thomas J. Watson, certainly impressed viewers and won the match, earning $1 million to be split between two charities, World Community Grid and apparently, oddly enough, the Canadian arm of World Vision. But given a large database and the capacity to navigate it quickly, it seems no great accomplishment to beat a couple of humans. Google could, if the person seaching it could only type faster.
What struck me was not Watson's success, but Watson's mistakes. (And is anybody else thrown by a "genius" who shares the name of Sherlock Holmes's goodhearted but not terribly perceptive sidekick?) Take one question -- I mean answer -- from Wednesday night's game, in the category "Nonfiction": "The New Yorker's 1959 review of this said in its brevity and clarity it is 'unlike most such manuals, a book as well as a tool." The correct response is The Elements of Style. But Watson bizarrely answered "Dorothy Parker." Parker is associated with the magazine, of course, though her heyday there was in the 1920s and 30s. But it's striking that a clue that clearly calls for the name of a book would elicit from Watson the name of an author. If he can make such a basic mistake, how can he ever hope to be one of our new overlords?
Watson's performance thus seems a vindication of a piece written beforehand by Maclean's magazine editor Colby Cosh. He makes the comparison, as many have, between Watson and Deep Blue, the IBM computer that took on -- and defeated -- chess grandmaster Garry Kasparov in 1997. But he asks how far we've come since then -- or, really, since the beginning of research into artificial intelligence:
Every article about Watson, IBM’s Jeopardy!-playing device,
should really lead off with the sentence “It’s the year 2011, for God’s
sake.” In the wondrous science-fiction future we occupy, even human
brains have instant broadband access to a staggeringly comprehensive
library of general knowledge. But the horrible natural-language skills
of a computer, even one with an essentially unlimited store of facts,
still compromise its function to the point of near-parity in a trivia
competition against unassisted humans. Surely this isn’t a triumph for artificial intelligence, or for IBM, so much as it is a self-administered black eye?
Colby asks an important question, one I haven't seen raised in all the hype about Watson's challenge:
So why, one might ask, are we still throwing computer power at such
tightly delimited tasks, ones that lie many layers of complexity below
what a human accomplishes in having a simple phone conversation?
Unlike the winsome Ken Jennings, I'm not ready to welcome our new computer overlords just yet.