“I, for one, welcome our new computer overlords.”
That’s what Ken Jennings, the famed champion of US trivia game show Jeopardy!, wrote alongside his final answer before losing to an IBM supercomputer named Watson in 2011.
It was a milestone for artificial intelligence, also known as machine learning or cognitive computing. IBM had been working on Watson for several years, hoping to prove to the world that it could do the impossible in a high-stakes media environment the same way it had done in 1997 with Deep Blue, a supercomputer that beat the chess world champion, Gary Kasparov.
To make Watson competitive, IBM needed to build a system that could understand linguistic nuance, scan a database of information it had previously ingested (Watson was not connected to the Internet), figure out a few possible answers and use statistical models to score them.
But other than putting human Jeopardy! champions in their place, what good was Watson? Back in 2011, IBM saw Watson as a possible tool for healthcare. Fast forward four years, and it is ready to unleash the power of Watson on a much broader scope.
“We recognized early on that there were areas of opportunity that IBM could get to and address, but a vast number of opportunities that we couldn’t attend to. So we opened up the platform,” Stephen Gold, IBM’s VP Partner Programs and VC Investments, told The Jerusalem Post at an IBM conference in Tel Aviv on Tuesday.
IBM is investing $100 million to push Watson into the business world and, according to The Wall Street Journal, IBM CEO Virginia Rometty sees Watson bringing in $10 billion in the decade through 2023.
The company has made about 25 different Watson APIs, software tools that developers can use, available on its website, and is seeking to build an entire ecosystem around Watson by letting entrepreneurs and developers figure out what to do with it – much the way Apple did with its iPhone app store.
“One of the big pushes in Israel today is doing just that. Israel has a truly entrepreneurial mindset – it’s a startup nation! – and we want to engage that nation of innovators and creators to bring forth new solutions, new offerings,” Gold said.
A tool like Watson is different from many others on the market precisely because it learns from its mistakes, taking in input on its performance, how its users are reacting to it and even on how often one source of information or another has proven reliable.
“There are a couple of dozen Israeli companies we’re already working with. It is amazing how quickly they have taken and adapted to this idea of cognitive computing,” Gold said.
At the conference, some of the amazing uses of the software were on display.
A University of Haifa-based IBM lab was using Watson to help radiologists identify irregularities in breast cancer scans, and quickly skim through diagnostic possibilities. It can even figure out that an ultrasound, for example, might be needed to narrow down the diagnosis possibilities presented by a mammogram.
Nearby, Israeli nutrition app Nutrino was showing how Watson could help its users get answers to tough nutrition questions by scouring a database of health articles. A Portuguese start-up, Novabase, was using Watson to help financial advisers find relevant information for their clients (e.g. what are analysts saying about a particular fund and are there better alternatives for a risk-averse investor).
All that builds on top of some of Watson’s growing list of accomplishments: It is helping Singapore residents do their taxes; researchers at the Baylor College of Medicine in Houston discover new cancer-related proteins; and even creating new fusion cuisines.
Believe it or not, Watson authored a cookbook’s worth of recipes, Cognitive Cooking with Chef Watson: Recipes for Innovation from IBM & the Institute of Culinary Education, simply by studying piles of other cookbooks (the recipe for Thai-Jewish chicken with potato latkes and rice balls sounds intriguing). I even enjoyed a sip of Watson’s Party Bourbon Punch at the event – refreshing and delicious.
Though Watson is not yet available in Hebrew, Gold hinted that could be coming down the line, as the team is developing languages “similar to Hebrew.” (Arabic was the fastest growing language on the Internet last year).
It is possible Watson will be coming to a smartphone near you soon enough.
Apple, Microsoft and Google have been in a battle of “personal assistants,” each trying to get better voice recognition, deeper contextual understanding and wider functionality.
Gold hinted that Apple’s Siri may be the beneficiary; Apple and IBM have collaborated of late on a host of enterprise projects.
“I fully expect that the kinds of things you can imagine will become very real in the months and years to come,” Gold said.
While the possibilities of utilizing a learning computer in a thousand different ways is scintillating, it also gives way to a Luddite pause: If Watson can do all this human stuff so much better than people, what will be left for people to do? Like Jennings, Gold is unphased.
“As long as there’s basic infrastructure, there are going to be jobs. No computing system has the capability to provide the physical manifestation of doing those jobs,” he said.
Not only that, but Watson needs to be “curated” for each of its uses, fed with the right data, set up to work properly, and get feedback from humans.
“There’s also the greater question of what’s the cost of not progressing,” Gold said.
If he’s wrong and Watson does end up taking everyone’s jobs, at least afford the human race will have more time to put up their feet and watch Jeopardy!