The other day, a classmate of mine brought up the question of whether Watson (the IBM computer who beat out our greatest champions of Jeopardy — because they were merely human) could become a replacement for reference librarians in my Emerging Web Technologies class the other day. Part of the concern was also due to an article in the New York Times about the computer intelligence of Google’s search algorithm, which also mentioned the Watson victory, in which stated,
“Google can be thought of as a supercharged, automated reference librarian for the Web.”
This comment seems primed to stoke the ire of the ruthless librarian/library-loving internet hive mind. (I’d rather not include any links to Steve Lohr’s contact information to facilitate the unleashing of its wrath — though I’m not sure that will stop it anyway.) But there is still the lingering question, and one that I’m more ambivalent about than offended by:
When will Watson be offering 24-hour reference service at my library?
For our class, the answer was left pretty uncertain – the discussion suggested that AI computers could only mimic human thought and not replicate it, but still that computers were growing “smart” enough in some important sense to allow them to respond to simple answer-based queries in Google’s Reference Tools: things like performing calculations and unit-conversions, and even suggesting specific answers to important questions like “What is the population of Florida?” or “How tall is Justin Bieber?”
To be sure, these are still pretty simple questions and don’t engage all the critical and emotional faculties which the librarian has to respond to library patrons confusion about how to get what they are looking for, helping them try to remembering/figure out what it is, or offer a sympathetic ear to the frustrated patron. These are important capacities, but I’m still not sure that this is reason to be skeptical about the idea that we could teach computers to think like humans or mimic human emotions to enough extent to handle much if not most of the queries that librarians receive at the reference desk. Though i suppose I can think of some reasons not to be excited about computer-automated reference…
Sure, I hate computer-automated telephone support as much as the next human being. But does this mean, as another classmate suggested, that we have a “puerile” sort of need to speak to a human librarian, even for the simplest questions? After the sight of Watson’s victory, it seems like a sufficiently human-like computer could do just fine. Or perhaps we can compromise on a pretty little replicant…?
P.S. For all who are interested in the discussion about Watson and artificial intelligence and whether computers can become like humans (or vice versa for that matter), there is an interview of Brian Christian with Jon Stewart on the The Daily Show: can’t figure out how to embed the video so the link is here. He is the author of The Most Human Human: What Talking with Computers Teaches Us About What It Means to Be Alive.
I haven’t read it, and I’m not sure Jon did either, though it still sounds pretty interesting. Anway, here’s a review from the Wall Street Journal by the British philosopher Julian Baggini, who referred to the book as “absorbing.” And here’s a post by Stanley Fish in the NY Times Opinionator blog called “What Did Watson the Computer Do?” and the follow-up by Sean Dorrence Kelly and Hubert Dreyfus called “Watson Still Can’t Think.”
P.P.S. I wish I could embed this much more entertaining video about Watson brought to you by IBM and CollegeHumor.com…alas, my WordPress skills fail me again.