Cheating is at the heart of the fear too many academics have about ChatGPT, and I’ve seen a lot of hand-wringing articles from MSM posted on Facebook and Twitter. One of the more provocative screeds on this I’ve seen lately was in the Chronicle of Higher Education, “ChatGPT is a Plagiarism Machine” by Joseph M. Keegin. In the nutshell, I think this guy is unhinged, but he’s also not alone.
Keegin claims he and his fellow graduate student instructors (he’s a PhD candidate in Philosophy at Tulane) are encountering loads of student work that “smelled strongly of AI generation,” and he and some of his peers have resorted to giving in-class handwritten tests and oral exams to stop the AI cheating. “But even then,” Keegin writes, “much of the work produced in class had a vague, airy, Wikipedia-lite quality that raised suspicions that students were memorizing and regurgitating the inaccurate answers generated by ChatGPT.”
(I cannot help but to recall one of the great lines from [the now problematically icky] Woody Allen in Annie Hall: “I was thrown out of college for cheating on a metaphysics exam; I looked into the soul of the boy sitting next to me.” But I digress.)
If Keegin is exaggerating in order to rattle readers and get some attention, then mission accomplished. But if he’s being sincere– that is, if he really believes his students are cheating everywhere on everything all the time and the way they’re cheating is by memorizing and then rewriting ChatGPT responses to Keegin’s in-class writing prompts– then these are the sort of delusions which should be discussed with a well-trained and experienced therapist. I’m not even kidding about that.
Now, I’m not saying that cheating is nothing to worry about at all, and if a student were to turn in whatever ChatGPT provided for a class assignment with no alterations, then a) yes, I think that’s cheating, but b) that’s the kind of cheating that’s easy to catch, and c) Google is a much more useful cheating tool for this kind of thing. Keegin is clearly wrong about ChatGPT being a “Plagiarism Machine” and I’ve written many many many different times about why I am certain of this. But what I am interested in here is what Keegin thinks does and doesn’t count as cheating.
The main argument he’s trying to make in this article is that administrators need to step in to stop this never ending-battle against the ChatGPT plagiarism. Universities should “devise a set of standards for identifying and responding to AI plagiarism. Consider simplifying the procedure for reporting academic-integrity issues; research AI-detection services and software, find one that works best for your institution, and make sure all paper-grading faculty have access and know how to use it.”
Keegin doesn’t define what he means by cheating (though he does give some examples that don’t actually seem like cheating to me), but I think we can figure it out by reading what he means by a “meaningful education.” He writes (I’ve added the emphasis) “A meaningful education demands doing work for oneself and owning the product of one’s labor, good or bad. The passing off of someone else’s work as one’s own has always been one of the greatest threats to the educational enterprise. The transformation of institutions of higher education into institutions of higher credentialism means that for many students, the only thing dissuading them from plagiarism or exam-copying is the threat of punishment.”
So, I think Keegin sees education as an activity where students labor alone at mastering the material delivered by the instructor. Knowledge is not something shared or communal, and it certainly isn’t created through interactions with others. Rather, students receive knowledge, do the work they are asked to do by the instructor, they do that work alone, and then students reproduce that knowledge investment provided by the instructor– with interest. So any work a student might do that involves anyone or anything else– other students, a tutor, a friend, a google search, and yes ChatGPT– is an opportunity for cheating.
More or less, this what Paulo Freire meant by the ineffective and unjust “banking model of education” which he wrote about over 50 years ago in Pedagogy of the Oppressed. Friere’s work remains very important in many fields specifically interested in pedagogy (including writing studies), and Pedagogy of the Oppressed is one of the most cited books in the social sciences. And yet, I think a lot of people in higher education– especially in STEM fields, business-oriented and other technical majors, and also in disciplines in the humanities that have not been particularly invested in pedagogy (philosophy, for example)– are okay with this system. These folks think education really is a lot like banking and “investing,” and they don’t see any problem with that metaphor. And if that’s your view of education, then getting help from anyone or anything that is not from the teacher is metaphorically like robbing a bank.
But I think it’s odd that Keegin is also upset with “credentialing” in higher education. That’s a common enough complaint, I suppose, especially when we talk about the problems with grading. But if we were to do away with degrees and grades as an indication of successful learning (or at least completion) and if we instead decided students should learn solely for the intrinsic value of learning, then why would it even matter if students cheated or not? That’d be completely their problem. (And btw, if universities did not offer credentials that have financial, social, and cultural value in the larger society, then universities would cease to exist– but that’s a different post).
Perhaps Keegin might say “I don’t have a problem with students seeking help from other people in the writing center or whatever. I have a problem with students seeking help from an AI.” I think that’s probably true with a lot of faculty. Even when professors have qualms about students getting a little too much help from a tutor, they still generally do see the value and usually encourage students to take advantage of support services, especially for students at the gen-ed levels.
But again, why is that different? If a student asks another human for help brainstorming a topic for an assignment, suggesting some ideas for research, creating an outline, suggesting some phrases to use, and/or helping out with proofreading, citation, and formatting, how is that not cheating when this help comes from a human but it is cheating when it comes from ChatGPT? And suppose a student instead turns to the internet and consults things like CliffsNotes, Wikipedia, Course Hero, other summaries and study guides, etc. etc.; is that cheating?
I could go on, but you get the idea. Again, I’m not saying that cheating in general and with ChatGPT in particular is nothing at all to worry about. And also to be fair to Keegin, he even admits “Some departments may choose to take a more optimistic approach to AI chatbots, insisting they can be helpful as a student research tool if used right.” But the more of these paranoid and shrill commentaries I read about “THE END” of writing assignments and how we have got to come up with harsh punishments for students so they stop using AI, the more I think these folks are just scared that they’re not going to be able to give students the same bullshitty non-teaching writing assignments that they’ve been doing for years.
I think all instructors should experiment with giving their writing prompts to a chatbot. If the bot gives an acceptable response, then align the rubric so that it’s not possible to get a passing grade by submitting the bland writing that bots submit. I’m not suggesting that we should require or expect that all students will submit bot-generated drafts, but if bots can easily generate coherent infodump paragraphs that contain a mixture of true, misinterpreted, and hallucinated details, then basic 21stC literacy skills should naturally include the ability to spot the kinds of falsehoods that permeate bot-generated arguments.
I think that’s exactly right, Dennis. That’d be a really good thing for folks to do before they keep giving the same prompts/assignments as before. Maybe they’ll have some of their fears alleviated in that while ChatGPT is impressive, it’s not nearly as powerful as the hype and paranoia in pieces like this one might suggest. Or maybe they’ll see how they need to finally update the bad assignments they’d been using for years.