The last time you saw Watson the supercomputer, he was probably cleaning up on the game show Jeopardy! But since roundly defeating reigning champions Ken Jennings and Brad Rutter, Watson hasn't been resting on his laurels. In fact, just two years after his debut, Watson - as Steve Gold, executive at Watson Solutions at IBM describes it - has been "going to work."
What sets Watson apart from the average, run-of-the-mill computer sitting in your cubicle (which, we would hazard to guess, has not recently won thousands of dollars on a television game show?) One, it can navigate the complexities of human speech - including complicated turns-of-phrase that stump other machines. Second, it can process enormous volumes of information, from to medical records to trivia about medieval literature. Third - and perhaps most impressively - it can learn, continually adapting and improving as it is taught and programmed.
"Independent of where it begins its education, it's continually getting smarter," says Gold. "With each iteration, with each outcome, with each new piece of information, it's progressively getting better at being able to interact and answer those queries and questions that are inputted."
Watson's unique and - dare we say it - human-like qualities make it an attractive tool for a variety of industries. Consider health care, for example. Imagine you enter a doctor's office for the first time and, with a few clicks of a mouse, your physician can look not only at your own medical records, but your family history, histories of people with similar symptoms, and the entire body of medical literature available. Today, when doctors make diagnoses, they're limited by the boundaries of their own knowledge and experience - a doctor in Alaska, for example, would not necessarily be able to immediately recognize the symptoms of, say, Lyme disease, which is found most commonly in New England. As a result, says Gold, about one in five diagnoses turn out to be incomplete or inaccurate. Watson can fill in those blanks.
That kind of accessible omniscience, however, can also have its drawbacks. Just as easily as it can help doctors make diagnoses, Watson can also - using the same information - tell insurance companies which patients will be the most expensive to insure. As Steve Gold notes, this friction between privacy and technology intensifies as technology becomes a bigger and bigger part of our lives. "The problem isn't specific to Watson - it's endemic of automation and technology at large," says Gold. "One can always argue that you can misuse and misappropriate technology in a way that is counter-productive."
To hear more about what Watson has been biding its time - including how the supercomputer could contribute to finance, education, law, and even call centers - tune in to the full interview above.