Wednesday, February 9, 2011

RE: Watson - IBM's "question-answering" machine (potential implications for libraries?)

Posting to NGC4LIB

Thomas Krichel wrote (concerning http://nyti.ms/g6J9Xe):
<snip>
B.G. Sloan writes
> What if we had sophisticated affordable "question-answering" machines in ten years? What would that mean for libraries?"
Why would that make any change? The idea that people go to see librarians to get questions answered is already many years out of date, isn't it? To get a question answered you look it up on the web or you ask an expert.
</snip>
This is absolutely correct: the future is with us now! The number of reference questions asked has tumbled and there is no reason to think that this will change anytime soon.

Of course, almost no question has a single "correct" answer, except for questions such as, how tall is Mt. Everest, or, Lincoln belonged to the Democratic Party--true or false? Almost every substantive question has several possible answers. For example, a question I was asked once pops into my mind: Does communism lead inevitably to Stalinism? Hard to answer with a yes or no! There is no single "correct" answer.

So, the traditional reference questions termed "ready-reference" are probably already gone from the reference desk. But questions that demand more thought and require a deeper understanding will (I hope!) always be asked and I don't see how a computer can answer those. The traditional library ideal that the librarian furnishes the searcher with information--in an unbiased manner--(or at least so far as is humanly possible) will still be needed, so that people can examine various ways of looking at an area of concern to them, and each can finally arrive at his or her own, personal version of "the truth".

How librarians can help people achieve this sort of ideal in a networked, virtual environment remains to be seen however, but this would seem to me to be one of the more interesting of the various challenges we face.

No comments:

Post a Comment