Recently, the Revolution was Televised (and I’m not talking about Egypt) – a computer built by engineers at IBM defeated the greatest champions that Jeopardy has ever seen – handily. Instead of predictions of the apocalypse, instead there has just been muted commentary on where the technology inside Watson will lead us in the near future – and what other more practical jobs (unless you’re a trivia expert, Jeopardy! probably isn’t that similar to your job) the technology can assist or replace. So, are you worried?
Deep Blue vs. Watson
Deep Blue was the original computer in the IBM Artificial Intelligence vs. Human Expertise grudge match. The first iteration of Deep Blue was handily defeated by World Champion Chess player Garry Kasparov in 1996 (It won once, then lost 3 times and drew twice). Taking on Kasparov in 1997, it beat him twice, lost once, then drew three times in a row. Kasparov then accused the IBM team of cheating, claimed he would never lose to the machine again… but had his rematch request rebuffed. Deep Blue had the ability to look forward 20 moves (‘plys’, in computer game programming) in some situations… but the age of computer dominance wasn’t yet at hand.
Even though early versions of Watson were only answering 15% of Jeopardy! clues correctly, adding
new algorithms to the software and a larger dataset to draw answers from allowed it to compete on a better basis. Jeopardy! staff made Watson move an electronic finger to buzz in during the grudge match recently competed – but with perfect reaction time, it was able to win the match with $77,147 vs. $24,000 for Ken Jennings and $21,600 for Brad Rutter.
What’s next for Watson?
Watson was a rousing success for IBM marketing, as well as an impressive demonstration of deep searching algorithms in huge data sets (Watson used sources like a downloaded version of Wikipedia as a search set). This leads us to our scary title – what sorts of things will Watson be good at once the software starts to be deployed to large corporate and government customers? Anything which requires analysis of images or text with a large number of supporting documents – think medicine and law – may be muscled in on by Watson’s technique. Also, Watson with a synthesized voice might be coming to call centers and other hubs of customer support. Watson also can be used to improve things like media recognition – think Google search, Google Scholar, Google Books, and Youtube… to name all Google Technologies.
However, even if some jobs on the margins are replaced by technologies like Watson, new jobs will be created… likely in high tech fields creating the next versions of software and even intelligence engines. Isn’t that what the concept of creative destruction is all about? To quote Ken Jennings, after his electronic defeat (quoting the Simpsons!), “I for one welcome our new computer overlords.” Do you?