A.I. In The Classroom

Mountain Connection, February 28th, 2023

By: Jed Donnel

With artificial intelligence all over the news lately, I am reminded of why I value teaching, particularly at a small, independent school. I do not share the various fears regarding A.I. that have apparently plagued many in the education field lately, and especially since the late fall when ChatGPT went live and became, by credible accounts, immensely popular nearly overnight among high school students who saw the app as means to produce knock-off writing they could submit as their own. Honestly, in my own experiments with the program, I found ChatGPT to lack credibility while feigning knowledge. at worst. Certainly, it can easily fool students – and teachers – who want a quick way through what should be a meaningful task. Even so, as an English teacher I think that the availability of such programming raises great opportunities. 

At the moment, ChatGPT produces writing that, while grammatically clean and reasonably fluid, lacks depth. It cannot assess quality, nor offer insightful critique (although it can generalize published, human assessment of quality in general terms). It cannot analyze nuance within citations, nor determine for itself the relative purpose of any citation that might be used to evidence a novel idea. What it can do is produce stock answers to factually-based questions, and, most of the time, it will do so using mostly correct information. In that sense, it’s not much different than lots of other information-based sites currently on the internet, including Wikipedia, and I am old enough to recall the general uproar among teachers when that site emerged, too.

It is therefore my role, as a teacher, to address such technology with my students, who are smart and engaging with a healthy sense of ethics and who, like all humans, appreciate being able to think for themselves. So, we’ve openly made A.I. a subject matter in my English classes, and we’ve begun to address what does and does not work within its intentions. Students have told me about their experiments with it, and I’ve displayed my own. Together, we’ve forecast how long it may take before the parent company of ChatGPT starts charging its users (not long, it turns out), whether the novelty of A.I. may wear off, and what may happen next. I’ve asked my students to find ways that such programs may actually save them time without compromising their integrity, to think of prompts that have not occurred to me, and to report back their findings without fear of repercussion. In fact, I’ve been quite impressed with what they’ve said to date, and certainly I’m curious to hear more. 

For example, I teach a senior English class on postmodernism, wherein my students are, on average, less optimistic about A.I. than national trends may suppose. Perhaps their reaction boils down to their natures, though I suspect that digesting the likes of Eggers and Althusser over the last few months has contributed, too. Recently, we’ve read two starkly different literary reactions to the idea of “The American Dream” during the Cold War era: Allen Ginsberg’s early poetry from the 50’s that waivers between fatigued cynicism and Whitmanesque egalitarian wonderment, and Joan Didion’s Play It As It Lays, one of the darkest novels written in the 1970’s that presents American ennui in beautifully stark prose. Rather than asking them to fully formulate their own considerations of the Dream in postmodern America, my students are first assessing the different presentations of the two authors. Then, to up the game and apply their literary critiques to topical consideration, they were recently  assigned to produce some of their essay using ChatGPT and to critique the results; they need to copy what the program spews back to them, paste it in blue font into their essay, and then take apart its qualities and shortcomings. What do they think of the technology as it applies to critical thinking? What are the current limitations of the program that may lead them to relate to Ginsberg’s critiques or (much worse) to Maria’s despair? On the other hand, what are its possibilities that may help us remember ourselves, in Ginsberg’s words, as sunflowers rather than locomotives? How might it help or hinder the American Dream? Most importantly, how does the availability of A.I. limit or enhance original thought? 

As I’ve emphasized, these are not trick questions, and I don’t know the answers. What I do know is that A.I. is not capable of offering interesting answers to such questions – answers with variety, with depth, with arguable positions rooted in detailed evidence. And I trust my students. I want to know what they really think and why. After all, they will lead the charge into the future, and their generation will determine the real and long-term uses of such technology. They may as well start now.

The Connection