Jeopardy!, Watson, and the Promise of Artificial Intelligence

Boy, that Watson was something else, wasn’t he?

I myself watch Jeopardy! regularly and even DVR it, so I’m sure you can understand my geeky salivating excitement at the computing experiment.  Even if you don’t usually watch Jeopardy!, I hope you had a chance to check out the anthropomorphized supercomputer plowing his way through Alex Trebek’s answers and questions.

Reaction to Watson’s win has been tinged with a bit of snickering and scoffing. That stupid computer thought Toronto was a U.S. city! His computer-quick reaction time gave him an unfair advantage! And so what if computers are good at trivia; they still can’t do anything good.

Despite the skepticism, Watson is a truly amazing feat of natural language processing and artificial intelligence.  The reason he triggers such reactions is in part, I believe, because of the accessibility of Artificial Intelligence.

Way back in 1956, the proposal that launched the field of AI (via a research summer at my alma mater Dartmouth College) had lofty goals even by today’s standards. The 10 scientists conjectured that “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” 55 years later, that reality still sounds just as plausible and just as futuristic.

Unlike other branches with computer science, the promise of AI is easy for us to imagine. Fifty years ago, few people dreamed of an Internet that would connect you with all the information in the world, mobile phones that fit in your pocket and do absolutely everything, or video games you control with your body.

But a machine that understands what you’re saying, cleans up after you, plays ping-pong? That’s easy to dream of, and hard to live up to. And Watson is one step closer.

Did you watch Watson on Jeopardy!? What do you think of the advances in AI that led to Watson’s development?