In the university right now, there is a paradox at the core of our project. Thanks to the pervasiveness of artificial intelligence, writing has suddenly become effectively obsolete and essentially impossible; it also has simultaneously become more necessary than ever.
This past semester was my first back in the rubble of freshman composition since GPT-4 bulldozed the pillars of what had been my writing composition class. For the past few years I had been teaching creative writing, blissfully ignoring most of the hazards of AI’s infiltration of the university. Returning to teaching composition, I found the situation to be as bad as I had feared.
Few pieces of student writing were beyond suspicion—and then usually only because of their very low quality. Sometimes I would use an AI to catch student AI use, sometimes I caught it myself, and sometimes I failed to catch it or didn’t bother trying. Students used AI not only for analysis and research in their essays, but for personal narrative and reflection, too.
The case for why students should have to take freshman writing has always been a challenging one to make to the students themselves.
I don’t blame them. In their shoes, I would have done the same. I didn’t want to write essays I didn’t want to write when I was a student. I still don’t. If, at 18 or 22, I’d had a way to produce better written work with almost no effort, I would have done so too.
The case for why students should have to take freshman writing has always been a challenging one to make to the students themselves. Often, we tell them writing is a foundational skill that can be applied in any major and any career. That’s true as far as it goes, but I never wanted to present students with such a reductive and prosaic view of writing. Instead, I told them that writing is a tool of thought, the primary practice by which we cultivate our minds; we don’t write to acquire skills that can be leveraged for success, we write to discover and create our very selves.
It was a self-consciously inspirational note I tried to strike. But how swiftly has AI relieved me of my best rhetorical maneuvers. Why do we need to think when there’s a machine that will do it for us? In the same way I can defeat five-time World Chess Champion Magnus Carlsen with access to a chess engine, a student with AI can think and write better than their professors, an employer, or anyone else in their life might require of them. Either they recognize the intrinsic rewards of writing as a profound tool of personal development (as I eventually did) and pursue it for that reason or they don’t and don’t. Either way, the trump card of “writing as thought” is off the teacher’s table.
A world that outsources writing to machines is a world that devalues human thought. Like most things most worth doing in this life, writing isn’t easy. Neither is teaching it to students who have no reason to do it at a university that doesn’t really care about education.
Am I too hard on my employer? As it stands by while professors use AI to grade work students produce with AI, I don’t think so. The tuition dollars keep coming in, the graduates keep going out. Isn’t that how this factory is meant to run? Never mind the decline in student effort and ability over the past several years. The world is still turning. Nothing to see here.
If I’m right that writing is obsolete and, for many, becoming impossible to master in the current environment, we have two options: either accept a world in which college graduates aren’t competent writers and thinkers or change the college environment.
So what exactly should we do? The Atlantic staff writer Tyler Austin Harper’s suggestions to ban artificial intelligence from student life are, at least for now, probably impractical. But what’s the alternative? Whatever we end up doing, we of the university must create the kind of environment in which our students no longer escape the responsibility of thought.
I don’t know what this means across the university. I know what it will mean in my classroom this fall: paper and pen, writing and revising, sentence by sentence, forever and ever, amen.
Fifteen years ago, one of my senior psychology research students designed an ingenious experiment. Using students in my two freshman composition courses, she applied something called the "CLOSE" technique. She replaced every 5th word in a 200-word essay with a blank. Students were asked to fill in the blanks. Grading was simply the percentage of correct responses for the 40 blanks. Subjects did this for 3 essays: an essay they had written, an essay they had read, and an essay they had never seen before. The success rates for the three types of essays were distinctly different: 90% correct for what they had written; 75% for what they had recently read and 60% for the new material (essays written by students the previous year). Four weeks later, performance on the essay they had written had not declined, but performance on the essay they had only read declined to the unread level. This demonstrated that the process of writing changes memory (and thus our brains) in ways that simply reading an essay doesn’t. And, by the way, any writing that students submit can be turned into a CLOSE test and show whether or not they actually wrote it. Also, the student researcher went on to earn a doctorate in quantitative psych and now leads the James Madison University Center for Assessment and Research.