The science section of yesterday’s New York Times had the headline—or at least some variation thereof— I’ve been expecting to see for a while now: New Test for Computers: Grading Essays at College Level. I still gagged a bit.
I’m not skeptical that a computer program with sophisticated enough AI software can grade basic essays. At this point, doing so is perhaps not even that great a challenge.
But why would we ever think this is a good thing? To me this development captures so many of the ills plaguing education, particularly an unrelenting push to standardize as much as possible.
Let’s consider the following example. I know it’s extreme, and it’s not the sort of work the computers would be grading. But bear with me to what I think will be a clear point. Students used to ask me frequently how long a paper had to be. Early in the year, I would bring in two of my favorite novels, Melville’s Moby-Dick and Hemingway’s The Old Man and the Sea, and juxtapose them. They are, I would explain, in many ways the same essential story told in very different ways. The kids grasped the message. I wonder how a computer would deal with either classic work.
That notion leads quite naturally into the same concept expanded. Language is tied to the same neurological expansion that enabled us to develop tools. It encompasses the higher of human capabilities, that amazing cerebral flexibility to merge the abstract and the concrete, to capture thought and imagination in ways that seem almost tangible to us. Not only that, but language also allows us to express ourselves in infinitely, highly individualistic fashion while unifying us as a community.
And at a time when creativity and communication are keys to solving the gigantic issues we face as a society—perhaps as a species—why would we willingly reduce the assessment of a vital human skill to a series of algorithms?