Am I too worried about Watson?

By | March 2, 2011
3 Comments

In my “Why Watson is Important” post, I asked, “If, instead of general knowledge questions, Watson were optimized to do computer support, teach a class, or practice law, how long until those sorts of jobs start to be done by computers?”

My answer was about 10 years.  The SciGuy blogger wanted to know if I could be right:

But is Watson really up for dispensing legal advice or customer support? Unsure of this myself, I turned to Vivek Sarkar, a computer whiz at Rice University who worked at IBM on the company’s Deep Blue project. I asked, “Is Watson going to take away our jobs?”

“I think that’s a little paranoid,” Sarkar said.

Essentially, Watson is analogous to a GPS system for navigating unstructured information, such as the mess that is the Web. The computer was endowed with the ability to process natural language — in this case questions — and browse the web for answers. It then ranked possible answers as the most likely.

While this ability to churn through the mass of information on the Web gave Watson an edge over its competitors, it was far from perfect, as its infamous answer of Toronto in a Final Jeopardy question (In the category of “U.S. Cities,” the question was: “Its largest airport is named for a World War II hero; its second largest, for a World War II battle.”)

Where Watson falls down, then, is its inability to separate the ludicrous answers from the plausible answers in its Web search. Therefore, Sarkar says:

“I think this is really a tool that can assist knowledge workers, but isn’t going to replace them outright. Just like we won’t let a GPS drive our car by itself, we’re not going to turn over legal advice to computer. It’s just not ready for prime time, and won’t be in 10 years.”

Sarkar’s thinking here is linear.  My GPS can’t drive my car, but there will be automatic automobile systems in 10 years, so his analogy gives small comfort.  Likewise, Watson will not remain frozen, only as capable in 10 years as it is today.  Watson and its descendent’s will continue to be refined and improved – both hardware and software.  


Computers and robots already take jobs.  Manufacturing jobs continue to be automated out of existence. There are fewer bank tellers per capita today than 20 years ago.  Nobody uses a travel agent anymore.  

Again, the computer does not have to be self-aware to take your job.  It just needs to do the job better or cheaper than you.  The SciGuy’s final point:

in a couple of decades we’ll probably all have a Watson equivalent in our pockets — or maybe embedded in our skulls, who knows — but it likely won’t have taken most of our jobs.

Either version of the future – where computers do most knowledge work and probably most physical work too – or where upgraded humans get the work… either is a pretty big deal.  Interesting times are ahead.


Related articles
Enhanced by Zemanta
  • dwwood76

    I think the question “Is Watson going to take our jobs” is a strawman.

    Of course Watson isn’t taking anybody’s job, unless that person’s job is to be a contestant on Jeopardy.

    Will Watson’s descendants take people’s jobs? Yes, absolutely. We may eventually have AI helpers like Watson installed in our craniums, but far too many jobs don’t require full human intelligence, as you’ve mentioned before.

    Why in the world would we upgrade humans to do work that could just as easily be accomplished by pure AI? Just to enjoy the drudgery of work?

    I think that we will have imbedded AI, and we will make extensive use of it, but that the vast majority of people will not be using it for work.

    I define work here as activity that is performed to pay for the needs and wants of life.

    Is a prosumer working? Is an unpaid blogger working? An upaid visual artist, songwriter, writer, social volunteer, etc. working?

    I think that what will change the most is our idea of what work is. Should people be compensated (at least enough to live simply off of) for contributing their thoughts, ideas and creations to the world? I think yes, since it enriches us all. At if it is the capital itself producing all the things we need to live, it is costing other human beings next to nothing.

    I believe that is where we will go and what we will use Watson’s descendants to help us do.

  • kjmikkel

    The New York times has a good article on the issue Armies of Expensive Lawyers, Replaced by Cheaper Software.

  • number1300

    This is a scary thought to ponder. If in the future these computer and future sophisticated programs would take over our jobs, what then are we going to do in order to make a living?