Assistive Technology

I woke this morning to the alarming news, via push-notification, from the NY Times that A.I. chatbots are forcing educators to rethink how we teach. The article explained that ChatGPT, the popular chatbot released by OpenAI in early December 2022, is able to answer questions and compose text in such an effective way as to pose a threat to traditional homework assignments. This bot can do math problems, research questions, and write essays all in simple and disarmingly-human sounding sentences. 

Suddenly eager to find out more about this technology which has been only on the periphery of my attention until now, I clicked to another recent article from the Times, which had 10 writing samples and challenged readers to identify them as either the work of a real student or a chatbot. The authors had asked a fourth-grade teacher; a professional writing tutor; a Stanford education professor; and Judy Blume, to see if they could tell which was which. None of them was right every time. I got 9 out of 10.

My next stop was the bot itself, which, not surprisingly with all its recent media attention was over capacity. I had to try several times before I could set up an account and start playing with it. The wait was absolutely worth it, though. When I finally got on, I tested several questions and prompts related to our current essay unit, trying to see what my students might find, if they knew where to look. 

The bot and I had some interesting exchanges about the characters in "Thank You M'am" and "Raymond's Run", but it was flat out wrong about more than a couple of its assertions about Squeaky. That made me feel better, and it was just fabricating ideas about the main character in the story "Vanquishing the Hungry Chinese Zombie". It misgendered her and spat out a bunch of hooey about saving her village, which was very similar to the patently wrong information it included in a claim about the character of Spark in "Bouncing the Grinning Goat". I literally laughed out loud at how wrong the dumb bot was.

So there was that quantum of solace, but as I thought more about my students, I focused on the one kid who won't write about any of the assigned stories for a variety of reasons, and I wondered if I could get the bot to tailor a story to his penchants. Well. It was a learning curve, and I required the bot to give me at least 15 revisions, none of which were perfect. I was able, however to stitch together a tale of 1500 words or so that has a character he may find relatable, and who accomplishes something he might find admirable, but with methods that are unquestionably unethical. That's a claim waiting to happen-- do the ends justify the means?

The writing? Is not remarkable, despite my instructions to add dialogue and figurative language, which the bot did. What the composition lacks is concrete details and present actions; it tells instead of shows, and I did not have the desire to revise for artistry. 

On the other hand, I cannot wait to see what this conundrum of a kid thinks of the story, which, by the way, is called "The Ends and the Means" by A. I. Chabeaux. Clearly, there will be a part 2 to this blog post, if not that story.