Tonight on Jeopardy, Watson, a computer programmed by IBM, will challenge the program’s two biggest human champions to see whether man or machine can dominate the field of recalling arcane facts about Potent Potables. Here’s an overview of how Watson and the competition will work; to understand the larger implications of artificial intelligence on our futures, you could do worse than to read my colleague Lev Grossman’s fascinating cover story on “The Singularity,” i.e., the notion that in the not-too-distant-future, machine intelligence will outstrip (and maybe merge with) human.
When I first heard about the challenge, I wondered what the big deal was: obviously a machine could easily be loaded with terabytes of trivia. The problem, though, is not getting a computer to offer the right “question” (in Jeopardy’s traditional inversion) but to understand the “answer,” which usually relies on wordplay, allusion and other non-literal language. In other words: the challenge is to get Watson not to recall facts but to reason, and thus know which facts to recall—the achievement that separates Wikipedia from, say, Cliff Claven.
I don’t know who wins, but I (and the Tuned In Jrs.) will be avidly watching. If Watson wins, I hope we mortals learn to appreciate the things that still distinguish us from our future Cylon masters. They’ll never build a computer that can defeat a human at Wipeout!