After trouncing its carbon-based competitors in Jeopardy, it seems Watson can answer nearly every question except one: "Where does Watson go from here?" To borrow a phrase from HAL 9000: I'm sorry, readers--I'm afraid it can't do that.
With this in mind, the company convened a group of experts yesterday at IBM's campus outside New York City. The aim was to provide "input on how the public and private sectors can harness technology for societal advancement," in order to "explore the ways in which IBM's Watson computing system might be applied to civic, social and cultural challenges."
Watson, of course, is IBM's new Question Answering computing system. Last week, Jeopardy aired a three-episode human vs. computer challenge to showcase the system. The match pitted Watson against the show's top all time human competitors--Ken Jennings and Brad Rutter. Watson mercilessly crushed the humans.
Watson is unique and groundbreaking for its ability to answer natural language questions by drawing on its store of natural language sources (e.g. encyclopedia entries and blog posts). Dr. David Ferrucci, Principal Investigator on the Watson project, explained that Watson is so remarkable because it can take a question on a broad and open range of topics, parse complex natural language and idioms, return precise responses, and evaluate the accuracy and confidence of its answers. Watson does this in a matter of seconds or less.
Watson presents new possibilities for private sector collaboration to solve social problems
Yesterday's invitees at IBM included leaders from myriad fields such as social services, health care, scientific research, environmental work, and disaster relief. IBM is wise to gather experts for their input, and also for beginning to explore ways to partner with nonprofits, NGOs, and researchers for meaningful impact.
Stanley S. Litow hosted the event as IBM's Vice President for Corporate Citizenship and Corporate Affairs, and President of the IBM International Foundation. He explained that the company's goal is to use technology, innovation, and creativity to solve business problems as well as social problems. In an interview, Litow elaborated that IBM hopes to develop collaborations with non-commercial partners. IBM brings the technology and resources; partners bring the field applications and expertise.
Such collaborations could be highly productive in addressing important educational and scientific purposes; IBM has a strong history here. For example, IBM developed its Reading Companion online literacy program by leveraging its expertise in voice recognition technology, working with an expert panel of top literacy teachers, and conducting data gathering in the field. In another instance, IBM worked with The Palace Museum to create "The Virtual Forbidden City," an immersive three-dimensional learning experience about Beijing's famous landmark. Watson's big Jeopardy win earned a $500,000 contribution to World Community Grid, another IBM partnership that uses grid computing to tackle large research problems.
Today's group imagined new possibilities for Watson. Guests from top hospitals discussed medical applications. Watson might, for instance, take a patient history, search the sea of medical literature, and present a doctor with a list of salient references and potential diagnoses. To Watson, no medical source would be too esoteric to mine for germane information. Watson might give TV's misanthropic Dr. House some diagnostic competition (and ironically, Watson might be better socially adjusted).
Educators in the audience buzzed about Watson's applications in the classroom. Watson could help teach, provide independent learning programs, prepare students for exams, and even help with the grading. Imagine a comprehension tool where students could read a passage or a book and freely ask Watson questions about it.
A relief worker pondered using Watson to support disaster response teams to learn about local infrastructure, culture, and medical problems. Others discussed using Watson on the frontline of social services to field everything from 311 calls to tax questions to health insurance inquiries.
When we consider the potential for miniaturization and portability, the possibilities seem truly boundless. Imagine a medical diagnosis tool or reading comprehension program deployed to isolated rural areas and developing nations. Just as the punch card systems and vacuum tubes of early computing are completely surpassed by the basic cell phones of today, one can imagine a time when Watson might be just as accessible.
Watson's performance reveals the value of uniquely human qualities
Of course, all of this exciting computer technology has significant implications for humans. As we envision and develop Watson's future role in the world, it's also important to envision and develop humans' future role in the world. Accordingly, many participants in today's forum expressed hope that Watson sparks new interest in math and science education; some speculated that engineering skills would be even more crucial in a Watson future.
Furthermore, if Watson-like technology is able to enhance or supplant so much research and analysis, we can put greater emphasis on improving our most vital human abilities. While Watson can dominate its Jeopardy competition, some of our most important qualities are our most uniquely human: leading, thinking strategically, forming interpersonal relationships, and making management decisions, to name a few. Watson can't lead a board meeting, develop a strategic plan, close a sale over dinner, repair a car, perform a ballet, sketch a portrait, or kiss a booboo. Nor can Watson imagine new creative endeavors and innovations--Watson could never imagine building Watson.
With a system to help answer the questions, what we do with those answers becomes even more important. A CEO might sit down and ask Watson for some information to support a decision making process. A salesperson might use Watson to find out some critical background about a contact, or learn about a new product. At the end of the day, though, a human will have to figure out what to do with that information. Investors won't back a company or a social enterprise or a philanthropic initiative simply because management is using a Question Answering tool. Investors will back a team because of their most human attributes.
Having Watson to do some of the informational and analytical work, people will be able to focus more time on creative, innovative, and interpersonal initiatives. For instance, if teachers can leverage Watson to help give lessons in remedial grammar, it might free up more time to teach creative and expository writing.
So where does Watson go from here? Watson can't answer that question, but it's a good thing there are a bunch of humans working to figure it out.
[With contributions from David Korngold]