That terrible “Everyone is cheating their way through college” essay
Recently New York Magazine published an article by James Walsh about AI usage in colleges. This article has a ton of traction among faculty who are frustrated with this moment and a political class who had already soured on higher education. It is also not very good. Every character is framed as the worst version of a college stereotype. There are some fundamental logical flaws in this analysis, but the larger problem is that it casts those of us in higher education as passive actors or non-playable-characters (NPCs as the kids say) stumbling through a world we don’t influence.
Bad Arguments
First, the logical problems. Walsh correctly notes “while professors may think they are good at detecting AI-generated writing, studies have found they’re actually not” and then (I’m not exaggerating here) three paragraphs later cites a professor’s intuitive approximation of AI usage as evidence for his larger point “By November, Williams estimated that at least half of his students were using AI to write their papers.” You cannot have it both ways. I do not understand how this sort of thing gets written, much less, approved by an editor. Maybe ChatGPT could have helped.
Second, early in the essay Walsh, again, correctly identifies that cheating is not new and in fact had been industrialized through Chegg and Course Hero. He then points to ChatGPT as the next evolution of academic dishonesty, but all of the hand wringing is about ChatGPT, not its predecessor. Nik and I have had conversations for the past two years about something I am finally going to say out loud–ChatGPT democratized cheating. In the academy we had basically made peace with the reality that essays and exams could be outsourced because it was mainly wealthy students who did it. Apparently our problem is not with cheating, it is with students who are poor being able to cheat.
Incredibly Obvious Oversights
The fundamental premise of this essay is that lots of students are using AI to write papers and do homework and so far we have not found ways of making sure they don’t do that. Except that we have. None of the solutions work all the time. One set of approaches involves strategies that try to get students to do the work themselves. At our home University we have developed a matrix where faculty map their existing assignments onto alternatives that are appropriate for different modalities of instruction. It isn’t magic, but it does require work.
The crudest solution mentioned in this article is students doing work in class. Lot’s of us were doing this already in flipped classroom models. It is not always elegant, but it does work for some classes.
Telling students why makes a difference. If we are being honest with ourselves, we got really lazy about this and outsourced accountability to plagiarism detection software instead of talking with students. The chemistry exam or the essay became an end unto itself and we stopped connecting the experience to a goal a student wants to achieve. If they understand why the process is more important they are more likely to engage it in good faith.
Question and answer helps. When I teach writing, students present their work and part of that is a robust question and answer following their talks. Maybe that is now a significant part of every assignment. Maybe the essay is now a poster session gallery where the presentation of ideas is more like an oral examination.
Unusual formats make a difference. I work with lots of faculty who have moved content production from just writing to also video, interviews, and podcast formats. Yes of course AI tools can aid in some of that work–it is a partial solution.
One set of solutions involves embracing the use of these tools for the tasks we ask students to do.
Jose Bowen argues, specifically in terms of writing, that we have to evolve our standards. If something was a C+ in 2021, it is an F now. Students can use these tools to create a baseline and we have to push them to do better than the stale formulaic writing many professors loathe.
I recently worked with a faculty member from another campus who had incorporated AI as a collaborator for his students at every stage of a multi-modal project. This is an example of what Ethan Mollick calls “Co-Intelligence” in his book by the same name.
This last point is more conceptual than specific, but if they are using these tools in the workplace then part of our job is teaching them how to do it responsibly and add value to what they are producing.
Pretending faculty and educators are not out there innovating is just lazy. No solution works for every class in every situation and we have to figure out what is the right fit for us, but the truth is–that was already the job.
The last fundamental misunderstanding in this piece is this idea that the University is an island. No part of this essay explored the world we are ostensibly preparing students to thrive in. When we talk with faculty one of the first things we tell them is “the world has changed, not just your class.” This is heartening in that it means we are in this together, but the ask is also much greater as we need to examine not just how we are teaching and testing, but also what we are teaching. The magnitude of this challenge is one reason we have been clamoring for help from accreditation groups for over a year. Walsh cites concerns about cognitive decline and critical thinking loss, but fails to connect the dots that it is the responsibility of educators to figure out how we can preserve or even enhance those characteristics in an AI empowered world. You could take a lot of this same evidence and retitle/reframe this article as “Why critical thinking matters in AI education,” but I’d guess that gets fewer reposts from anxious professors and angry politicians.
I apologize if this is more of a screed than we usually write in this space, but this article was optimized for academic doomerism and cynicism about education and I have no patience for either. This is a moment of disruption and overwhelm but we don’t have to be passive actors in this drama. We can figure this out together, but we need to actually do the work, not just complain and throw our hands up in the air.