Last night on Jeopardy, a machine beat two humans in a game that requires the kind of contextual understanding once thought impossible for computers. IBM's supercomputer Watson beat two of Jeopardy's all-time greatest champions, Ken Jennings and Brad Rutter. How significant of a milestone is this? In his final Jeopardy answer, Jennings jokingly wrote "I for one welcome our new computer overlords." That might be jumping the gun, but what does this historic Jeopardy victory say about the future of innovation?
For those of you who don't read technology blogs or plan your evenings around Jeopardy, let me quickly fill you in on how Watson works. What makes Watson a breakthrough is the ability to understand the question Alex Trebeck is asking, with all its allusions, puns, and witticisms. Unlike a search engine, which can point a human user to documents which may contain the answer, Watson searches for all possible answers and predicts which one is the right one. Watson calculated a confidence level for each answering, which refelcts how likely it is to be correct.
Watson is not always right. In final Jeopardy, when the question asked which American city has one airport named after a WWII hero and a second named after a WWII battle, Watson's answer was "Toronto????" Most humans will recognize that Toronto is not an American city, but I admit that issue slipped past me as I watched because I was focused on the curious question marks. The question marks indicate Watson's lack of confidence in the answer. One wonders if, when Watson is extremely confident, he puts "!!!!" or draws obscene messages like Darrell Hammond's characterture of Sean Connery on SNL's Celebrity Jeopardy spoof.
Watson's feat is reminiscent of another IBM innovation, Deep Blue, which famously defeated grand chessmaster Garry Kasparov. As this NY Magazine profile of Watson points out, Deep Blue generated lots of publicity for IBM but it never led to a business opportunity. The folks at IBM believe that computers like Watson will have plenty of practical uses, possibly in the healthcare and goverment industries. It is not hard to imagine how a technology with the computing power like Watson's could revolution revenue forecasting and risk management (In double and final Jeopardy, Watson curiously bid specific amounts down to the exact dollar in double Jeopardy because the team at IBM had programmed him to determine mathematically determine the correct bet). But I would like to suggest another aspect of innovation in which Watson would be a star right now, in its (his?) current form: creative problem solving.
Imagine Watson in a brainstorming meeting. Someone kicks off the meeting by stating the task or the problem the group is trying to solve. Watson understands the question and immediately starts spitting out relevant information, analogous historical examples, and possible solutions.
What's more, even when he was wrong he would be helpful. In brainstorming meetings, one idea often triggers several others. Watson would be the ultimate idea stimulator, providing an endless stream of fodder for human colleagues to play with. Even his most amusingly off-topic suggestions could provide humor for the group and lead them to think of something valuable.
Finally, Watson would never reveal his dislike for anyone else's idea. In all our meetings at Creative Realities, we "ban the bazooka," preventing the shooting down of anyone's ideas. Sometimes people shoot down ideas intentionally and verbally, but sometimes it happens as a subconscious or non-verbal reaction. Watson would never do that; he would always remain perfectly neutral.
Then again, he wouldn't be great at building on anyone else's idea, contributing positive energy, or committing to follow up on the idea...
I guess, for now, we still have the upper hand over our computer overloads.
By the way, you can play against Watson yourself at the New York Times multimedia game. I started out strong, but Watson ultimately beat me soundly thanks to the "musical pastiche" category. I still have no idea what music has to do with a delicious, shelled nut; but Watson knew.