Writing recently at The Chronicle of Higher Education, Corey Robin, professor at Brooklyn College and CUNY, as well as a high-profile author and public thinker, offers a stirring description and defense of what it really means to write.
He says, “It [writing] requires you to make your fleeting thought a hard fact in the world and to make yourself answerable for that fact. If you can’t do that, you need to revise your thought and find another fact, one that you can live with as your own.”
Here he well articulates one of my mantras about writing: writing is thinking. Writing is both the expression and the exploration of an idea. We must consider how our ideas are best expressed to an audience and engage with the fact that the idea itself may shift, even as we are engaging in this expression.
This is a very difficult thing, as Robin notes. Writing “doesn’t come naturally”; it is “effortful,” “frustrating” and “disappointing.”
“Failure looms large,” he says, invoking another core part of what I call “the writer’s practice,” the notion that failure—measured against our own intent for the clarity and impact of our writing—is literally inevitable.
He goes on to extol the virtues of “discomfort” as part of the experience: “Our goal shouldn’t be to eliminate this discomfort,” he says. We need to teach students that it’s part of the process and develop strategies for coping with it. But for students to really get that—to believe it—to feel it—they have to do the work.”
Yes, exactly. Amen. I’ve written in the past about the necessity of making students “uncomfortable” in order to foster a good atmosphere for learning. Having that sense of exploring unknown territory creates an exciting mix of anxiety and possibility and working through those feelings often results in students increasing their senses of agency and self-efficacy.
In my view, there’s no better way to achieve this than asking students to write.
Given the clarity with which Robin sees the importance and challenges of writing, it’s disappointing, then, that he concludes his essay by saying that this semester, for the first time in his 30-year career, “Instead of take-home essays, I’ll be requiring in-class writing, including midterms and finals.”
Even those who haven’t read Robin’s essay can probably guess the reason why he’s going this direction: GPT-4. Having been shown what kind of outputs can be coached out of the large language model through the skillful use of prompts and how closely those products resemble what he is looking for in his students’ writing, he feels the integrity of the process he has followed over his 30 years of teaching is inevitably compromised, because he will not be able to tell if his students have done the work, and if they don’t do the work, what’s the point?
I am sympathetic to these concerns. Generative AI large language models like ChatGPT and its more powerful cousin, GPT-4, are undoubtedly disruptive. What has been done in the past may not be possible to do in the future with the same sense of security and integrity.
But reading Robin’s lament, even as I couldn’t agree more with his core framing of what writing is and why it’s important, I think he is allowing the disruption to call the tune, rather than taking a step or two back and considering how the pedagogical experiences he values can be maintained or even enhanced in an AI world.
I think there’s three main issues worth addressing here:
- The (false) notion that what has gone on before represents an ideal that must be maintained or recaptured.
- The (false) equating of grades with “stakes.”
- The difference between an academic activity and a learning experience.
Robin believes that the process students go through in producing writing for his class—multiple drafts, intensive comments on drafts, revision and conversation—demonstrates to students that “all writing is rewriting and good work is just that: work.”
In theory, all of this is true. In practice, we know that this is not necessarily and certainly not always the case. Differences in student abilities can sometimes skew the ideal. I’ve had numerous students who arrived with sufficient skills to get A’s on their class essays without much sweat, turning these worthwhile exercises into pro-forma hoop jumping. The only thing they learned is that they already knew how to do something well enough to get the grade they desired.
Other students often choose to go through the motions of these worthy exercises, accepting their substandard grades as a consequence. Still others simply cheat.
In other cases, the ideal Robin articulates was not possible even prior to GPT. In my own career as a college instructor, the sequence Robin has employed—multiple drafts with extensive instructor comments—was never an option because I was working with far too many students in a semester to engage in those practices.
We cannot claim that ChatGPT has ruined a utopia if the utopia never existed.
I don’t think Robin would actually claim that the previous status quo was a utopia, but his essay makes clear that he views the instructor’s authority to grade as a backstop against students either avoiding the important work or being appropriately punished if they do avoid the work. In his view, the potential cudgel of that bad grade provides the motivation to dig in and do the occasionally unpleasant work of writing.
This may be true for some students, but it is not as true for most students, as many believe, and it is definitely not true for the writing we do outside of school contexts. True, meaningful stakes for writing attach when the writer cares about the impact of what they have to say, when what they’re writing matters—to the author most of all.
It’s best if there are real stakes attached to the work, for example, an authentic audience the student is writing to. A subject on which students have both sufficient interest and knowledge in order to feel as though they can write convincingly to this audience also matters a lot. Having the time and support to explore the challenge and go through the inevitable failures and frustrations that attach to writing is also necessary. In-class writing has a role to play, but it is a different role than the virtuous practice of writing that Robin extols.
I sometimes think that working under significant systemic constraints, which made teaching in the way Robin has been allowed impossible, forced me to become a pedagogy problem-solver, to disentangle the academic activity from the learning experience. If I simply cannot do extensive comments on every student’s draft, what can I do that provides the opportunity for them to learn?
These constraints made me realize that much of what I was doing in terms of student activities was actually divorced from the learning I felt was most important.
For years I assigned the kind of researched essay that anyone would recognize, an exercise I had absorbed as important from the teaching folklore I’d been immersed in, but which, in practice, inevitably resulted in disappointment for both me and the students themselves. Creating an imitation of an academic artifact was not helping my students learn to think and struggle like scholars.
ChatGPT/GPT-4 may be new, but the challenge of getting students engaged with learning is eternal. To that end, rather than focusing my efforts on how to maintain the integrity of an existing activity, e.g., a research paper, I recommend stepping back and rethinking activities around “learning-oriented” questions:
- What do we want students to know?
- What do we want students to be able to do?
- What attitudes do we want to foster in students?
- What are the experiences that will help students engage with these questions?
What this looks like in terms of ultimate assignment design will vary greatly, depending on the answers to these questions. In my asynchronous online course on teaching writing in an AI world, I work through the process that helps people uncover these answers and then design assessments that match one’s pedagogical values.
Maybe this is the “end of the take-home essays,” as Robin questions in the title to his piece, but if so, we must be able to put something worth doing in its place, no?
There’s a central irony to Robin’s concerns in that he both fears GPT-4 because students can use it to get the grade they desire, while also maintaining that the best (or perhaps only) way to make sure students are doing the substantive work is through the power of the grade.
This is not true in my experience, but even if it is true, rather than capitulating, shouldn’t we channel the spirit that Robin convincingly argues is central to the importance of writing—doing the hard thing, risking failure and trying again?
I think the integrity Robin posits for the pre-GPT world is illusory, but even if it isn’t, we can’t roll back the clock to last year, when this technology didn’t exist.
Writing is challenging and frustrating, but it is also awesome and empowering. Those are the experiences students should have with writing. To the extent that GPT makes us look harder at how those things happen in a school context, I’m pleased that this challenge has arrived.
I think few of us who know the power of writing believe that GPT has truly obviated the need for humans to write. If that’s true, we should make an even greater effort to make sure that students have the kinds of experiences that expose them to the pleasures and pains of writing.
The same spirit that animates the challenges of writing also animates the challenges of teaching.
Just like instructors must give students the resources necessary to meet the challenges of learning, it is on institutions to provide the time, space and expertise to faculty to work through these fresh challenges, something that I fear is happening only in certain pockets of higher education. There is lots of expertise about how to address these challenges available. I know I’m easy to find.
But once those conditions are met, it’s on faculty to be bold and discover what works in this world, rather than retreating to safety.