Tuesday, November 01, 2005

Computer Scientists Discover the Brandeis Model

BBC reports that two guys at Cambridge have come up with an AI system that automatically identifies and answers FAQs. Sony plans use it to provide tech support for Play Station 2 online. Questions that can't be answered by the system will be forwarded (by e-mail) to a human who will answer the question and add it to the MetaFAQ.

Look for more chicken little reports of how Reference Librarians are about to be replaced by machines. Take a deep breath. Do not be alarmed.

Play Station 2, a human designed computer system that is sold to customers on the assumption that it works. One would hope that most of the technical support questions would already be easily identifiable and aswerable. If that's what you are doing in the reference office then shame on you.

This work replicates Virginia's theory of reference service that's more than a decade old--and very misleadingly called the Brandeis Model. With the Brandeis model we use graduate students instead of a MetaFAQs to answer the easy, repetitive questions and human beings (not e-mail msgs) are passed back to the reference librarian when a question falls outside those limits.

Reference librarians in the RCO don't really answer questions. What they do is interview patrons to figure out what they need. More often than not, they can't articulate their research needs. A student might walk up to the desk and ask where the microform machines were. A half hour later we could easily be knee deep in the guide to ESTC on microfilm trying to find a pamphlet her professor mentioned in class.


My advice to people who are worried about Artificial Intelligence machines taking over human intelligence jobs is to read a bit in macro-linguistics, the philosophy of language and throw in a basic reader on the social construction of reality.

Just make sure you have a job that calls for creativity and human conversation and you're safe.
http://news.bbc.co.uk/2/hi/science/nature/2055491.stm

No comments: