EDITORIAL

The Watsons Are Coming

Hey, hey, we’re the Monkees,
You never know where we’ll be found.
so you’d better get ready,
We may be comin’ to your town.

—Theme from The Monkees
By Tommy Boyce and Bobby Hart

I n the latest iteration, it’s the Watsons who are coming … thousands at first, millions and billions soon to follow … and they truly are the “young generation, and [they’ve] got something to say” (yet another line from the Theme) … something so important that it will change our lives … forever.

Perhaps you read the recent news story about IBM’s Watson supercomputer beating a group of contestants in the TV game show Jeopardy! In a technological display of genius, Watson has achieved a milestone. This means that artificial intelligence (AI) is developing much more rapidly than most people realize. In the TV showdown, two of Jeopardy!’s greatest champions were matched against Watson, and the new generation of supercomputers won by a substantial margin. This is a clear demonstration that supercomputers can compete at levels that makes it far harder to argue that there are human tasks that computers will never realize.

The March of AI

Watson is an AI computer system with the capacity to answer questions posed in natural language, growing out of IBM’s DeepQA project, and named for IBM’s first president, Thomas J. Watson. In the instance of the Jeopardy! match, Watson demonstrated that it could understand great complexities involving metaphors, analogies, puns, humor, and ironies among other subtleties. At the same time, AI computing is on the march elsewhere with computers moving ahead on many other fronts, from cars that drive themselves to the diagnosis of disease. Wow!

When Will Watson Come to Your Town?

It is just a matter of time for Watson to arrive as a mass market phenomenon, putting this capability into your hands, meaning your computer or—better yet—your mobile device.

As we hear from author, inventor, and futurist Ray Kurzweil, the computer price to performance ratio is now doubling in less than a year, “so 90 servers [the currents number driving Watson] would become the equivalent of one server in about seven years, and the equivalent of one personal computer within a decade. However, with the growth in cloud computing—in which supercomputer capability is increasingly available to anyone via the Internet—Watson-like capability will actually be available to you much sooner.” 1

Kurzweil estimates that this “natural language processing” (the ability to “understand” ordinary English) will show up in the popular search engines over the next five years. This will soon lead to computers passing the famous “Turing test,” in which “chatterbot” programs are able to fool experts into believing that they are human. Ironically, Watson might have to dumb itself down to pass a Turing test. It would be too smart.

But is there such a thing as too smart when it comes to life and death matters, such as in the biomedical sciences and arts? Why stop at machine intelligence that is merely the equivalent of biological human intelligence? Why not thousands or even millions of times smarter than us? However, the real question to ask is: why can’t we incorporate future Watsons into ourselves to immensely expand and extend our own intelligence by merging with the tools of supercomputers. After all, they are our own creations.

Watson BioApp

IBM has just announced that they are already moving into the healthcare arena to commercialize Watson-like technology, in alliance with Nuance Communications. 2 The two companies expect the first commercial offerings from the collaboration to be available in 18–24 months.

For example, a Watson-like biomedical application will find use in indexing the medical literature, and data mining from free text, having the ability to discern between “what researchers mean” versus “what they wrote” and “what is consistent with all the evidence.” This alone has enormous implications for the healthcare sector. Patients will start to use the terms doctors learned in medical school to describe their ailments and thus learn how to translate that information or use it to elicit more information as needed in order to provide care.

Importantly, Watson BioApp would untangle the numerous conflicts of interest and dishonesty-related phenomena appearing in the medical literature. Plausibility issues would also fade, with Watson continuously on the case … applying scientific standards with greater precision, ironing out the contradictions, resolving what is known from what is not known.

Massachusetts General Turing Test

As Watson BioApp develops exponentially (or with the advent of quantum computing, factorially), the real breakthrough will occur when it can take, for example, cases from the Case Records of the Massachusetts General Hospital in the New England Journal of Medicine verbatim, and compete as a peer (not recognized to be a machine) with a round-table panel of expert physicians with access to PubMed, or equivalence, and go head to head with the supercomputer on the differential diagnosis, as well as how to establish the diagnosis and to rule out others. Furthermore, Watson BioApp will outcompete the peer panel in determining the treatment strategies, and the likely outcomes. And finally, Watson BioApp will direct the care. When this happens—let’s call it the Massachusetts General Turing Test—on an affordable level, the real biomedical revolution will be underway.

References

  1. Kurzweil R. When computers beat humans on Jeopardy.Artificial intelligence is developing much more rapidly than most of us realize. The Wall Street Journal, February 17, 2011.
  2. IBM Inc. (2011) IBM to collaborate with Nuance to apply IBM’s “Watson” analytics technology to healthcare [Press release]. Retrieved from http://www-03.ibm.com/press/us/en/pressrelease/33726.wss, February 17, 2011.

FREE Subscription

  • You're just getting started! We have published thousands of scientific health articles. Stay updated and maintain your health.

    It's free to your e-mail inbox and you can unsubscribe at any time.
    Loading Indicator