“Cheating is now so rampant that uni degrees have become worthless”
Plus, another op-ed from Middlebury College. Plus, Best/Worst is coming up again. Plus, an integrity award.
Issue 328
Subscribe below to join 4,216 (+6) other smart people who get “The Cheat Sheet.” New Issues every Tuesday and Thursday.
If you enjoy “The Cheat Sheet,” please consider joining the 16 amazing people who are chipping in a few bucks via Patreon. Or joining the 44 (-1) outstanding citizens who are now paid subscribers. Paid subscriptions start at $8 a month or $80 a year. Thank you!
“Cheating is now so rampant that uni degrees have become worthless”
That’s the headline of an opinion submission (subscription likely required) in the Sydney Morning Herald by Dr. Mindy MacLeod:
a former lecturer who now tutors at the University of Melbourne.
To start, I love that an opinion piece on cheating can get space in a major national daily paper. But it’s Australia, where people are less inclined to pretend that cheating is not a major scandal.
The MacLeod piece begins with her experience receiving “weird” e-mail messages from students:
It’s pretty easy to spot the ones generated by ChatGPT. When a student who can barely speak a word of English tells me “your advice is invaluable for focusing my efforts effectively”, it’s not hard to understand they didn’t come up with that on their own. No human actually would.
Nor would any normal student tell me they are “open to any recommendations you may have and am willing to adjust my focus based on your advice”. I am glad that my “guidance and expertise are invaluable” and that students “look forward to receiving [my] response”.
Later, she writes:
Rather disappointingly, my first written assessment of 2024 identified that at least 52 per cent of students had used AI to complete some or all of their papers. That’s probably an underestimation by the university plagiarism detector and is, in any case, a huge increase from the second semester of 2023.
I have to admit, I am surprised. That it was only 52%. And I do agree, that’s probably an underestimation.
She continues:
By the end of the second semester, even the AI detectors had given up, and only the laziest students were getting caught. The “smarter” ones were obviously using AI, but editing it enough to be unprovable.
No surprise here either. Students talk to one another and they share tips for cheating and tips for getting away with cheating. If our author addressed the first wave of cheating with her students — the 52% — the rate should fall thereafter because some students won’t cheat for fear of detection, others will invest more in avoiding detection, and others will find different ways to cheat.
The piece continues:
It isn’t just a matter of using AI to generate ideas, as the optimists would have you believe, or levelling the playing field between domestic and foreign students (as some idealistic academics would have it) by cleaning up grammar – it’s a matter of students telling AI to write a paper on the very subject they can’t be bothered to. Or, at the very least, an introduction. Or a conclusion. Or the methodology section. But more often than not, the whole damn thing.
Preach.
She goes on:
The university as we know it is dead. COVID-19 nearly killed it, taking everything online and out of control, and AI has dealt it its death blow. Until we alter the way we assess students, a degree in many faculties is a waste of money in terms of everything but earning (rather than learning) potential.
Concluding:
Detection technology will keep improving, of course. But so will AI. Until we can be assured that students are given opportunities to fail and that their degrees prove they know something about the subjects they are supposedly specialists in, they’re as meaningless as the essays students are submitting to swindle their way through their courses.
I have nothing further to add.
Another Op-Ed From Middlebury College
Middlebury College has found itself in the epicenter of discussion and action regarding academic integrity at capital-H Honor Code colleges (see Issue 322 or Issue 204 for examples). I believe what is happening there, or what will happen there, is important — whether a school, having acknowledged the deficiencies of its Honor Code, will also find the courage to act.
With that backdrop, the student paper at Middlebury has published another op-ed on the topic. Though it does not say, it appears to be from a student.
The piece starts well enough:
Anyone paying attention to higher education can clearly see that institutions have a cheating problem. Cheating first exploded during the Covid-19 pandemic and has only continued to rise with the dawn of generative artificial intelligence. This has led to a moment of reckoning for many institutions: How do we reconcile our traditional honor codes with the modern reality of widespread cheating?
But it’s a train wreck thereafter.
Before getting to the ignorance, I share this from the piece:
At the end of my first year, my psychology professor caught an organized cheating ring in his senior-level seminar. He then sought departmental approval to proctor exams for all his courses and has not looked back since. Incidents like these are a tragedy for the Middlebury community.
Fair. And, to underscore, Middlebury — like every other school in the world — has a cheating problem.
But then we go off the map:
You may think this will be an op-ed calling for stiffer punishments, blanket permission to proctor exams and the construction of an academic panopticon on Battell Beach. However, I do not think such sweeping measures are a productive or valid response to increased cheating or the shortcomings of the college’s Honor Code.
Our author instead suggests:
We should reinstate signing the Honor Code during orientation to establish its importance for each incoming Middlebury class. Studies investigating cheating behaviors have found that students learn how to cheat from their peers, which emphasizes the importance of early intervention. The signing could be coupled with workshops that engage students in discussions about academic integrity and why it matters, framing the Honor Code as a shared commitment rather than a passive agreement.
I rightly mocked this “solution” in Issue 322. Will it do any harm? No. Will it address cheating in a discernable way? No.
Also, as told to me once, no one cheats in college for the first time. Maybe cheating is learned and maybe it’s learned from other students. But that learning and practice takes place in middle school. By college, it’s too often too late. That same someone also told me that students show up to college swearing off their misconduct habits, pledging to do it right this time. But the first time that opportunity meets justification, the resolve dissolves.
Anyway, our writer continues with his vision for addressing the “cheating problem” that he calls “a tragedy”:
Considering the pace at which tools like ChatGPT are developing, and the near impossibility of accurately detecting their use, our efforts are better spent modifying the environment and behavior of the students that use AI. Faculty should design assessments that are less conducive to cheating, such as open-book exams, sit-down written essays or assignments that require critical thinking and unique applications of knowledge.
I left in the one link in case you want to check my math.
The link to “near impossibility” is not research. It’s a link to advice to teachers on dealing with AI use. Its only evidence that AI detection does not work is a single news article in the Washington Post from April of 2023, just days after the first AI detection software was released. A reporter tested it with unscientific rigor and said it did not work to his satisfaction. That’s it.
Since then, the actual research is uncontroverted — good AI detection systems work really well at detecting text created by AI with contrary “false positive” error rates of just 1%. To simply brush past the phrase “the near impossibility of accurately detecting [AI] use” is wrong.
And open-book exams are less conducive to cheating? Since when? Oh, right — since never. Unsupervised, open-book assessments are so easy to cheat they carry nearly no measurement value whatsoever. And assessments that require critical thinking “and unique applications of knowledge.” Why did no one else think of this? Oh, right — because that does not work either.
The article continues:
When violations of the Honor Code do occur, our goal should not be to shame and expel. Instead, we should focus on educating and reforming. Reflection essays and community service tied to academic workshops would educate rather than purely punish.
I know — or at least infer — that the author is a student. So, I apologize. But this is nonsense. When students understand that the consequences for academic fraud are to write an essay or forced “service,” cheating will likely increase. If the bookend outcomes are failing a course versus writing a reflective essay — come on.
He goes on:
While studies have found compelling evidence that close monitoring strategies can eliminate cheating, we should measure these interventions against the ideals of the Honor Code. Yes, the scrutiny of more proctors for in-person exams or lockdown browsers for online exams would probably significantly reduce cheating. But
I stopped at the word “but.” If you care about integrity and fairness, nothing that follows that word matters. If you’re balancing cheating against some other value or belief, you accept that there are conditions under which cheating, or allowing cheating, is acceptable.
Research has repeatedly shown that “close monitoring” of exams and other assessments reduces cheating. There endeth the sentence.
For the record, after “but” is this:
these measures would alienate us from the spirit of the Honor Code.
You mean the Honor Code that’s not working? The one that students are ignoring? Close monitoring of tests would shred the spirit of that Honor Code? Not the cheating? Even though it would work, it’s the close monitoring that would violate the spirit of the Honor Code. Got it.
Finally, disconnected from nearly everything else in this article, the author shares:
While institutions like Middlebury must foster academic integrity, students are not blameless. Cheating is a choice, and upholding the Honor Code starts with individual accountability.
Students are not blameless — big admission. And he wants accountability, though he does not favor punishments or integrity safeguards such as exam proctoring.
Finally, the author says he wants:
a culture where integrity is not enforced but embraced.
Sure. Thing is, we are having this conversation at Middlebury because we tried that already.
Best/Worst 2024 Nominations, Integrity Award
It’s that time again. Well, nearly.
If you’re a new reader of The Cheat Sheet, around the first of the year I do a “Review of the Year” Issue — see Issue 2021, Issue NY22/23, or Issue NY23/24 for examples.
If you’d like to suggest someone or something for one of the review categories — quote of the year, person of the year, story of the year, worst something, etc — please send them along. A reply e-mail to The Cheat Sheet reaches me.
Unrelated, last year was the inaugural awards program at EdTech Chronicle and, well, it’s open again. The award categories are in the submission form and, like the first year, the awards focus on people and what they do rather than on products. I mention these awards because there’s an integrity category — Best Product or Service Supporting Academic Integrity. So, if you’d like to apply for that one, or others, or know someone who should, please hit the link just above.