Buzzfeed Tries to Write about "False" Cheating Allegations. It's Very Funny. And Very Sad.
Plus, a funded PhD research position on academic integrity. Plus, Portland's possible non-zero for cheating.
Issue 245
To join the 3,656 smart people who subscribe to “The Cheat Sheet,” enter your e-mail address below. New Issues every Tuesday and Thursday. It’s free:
If you enjoy “The Cheat Sheet,” please consider joining the 14 amazing people who are chipping in a few bucks via Patreon. Or joining the 24 outstanding citizens who are now paid subscribers. Thank you!
Fun with “Wrongly Accused” of Cheating
At times, I think “The Cheat Sheet” is a bit too serious. I take integrity and the value of education seriously and I admit, sometimes things really tick me off.
You’re shocked, I know.
So, I’ve been holding on to this gem since August and it feels like the right time to get into it. It’s an article written over at BuzzFeed but syndicated globally - here it is at Yahoo!, for example. And since the article was clearly designed to be shared on social media, it was. Widely.
The headline is:
I'm Glad I'm Not In School, Because These 12 Teachers Who Wrongly Accused Their Students Of Cheating Are Infuriating
To start, it’s written by the same person who penned other insightful pieces for BuzzFeed such as:
16 Really Bizarre Happenings That People Got Photo Evidence Of — I Wouldn't Believe It Otherwise
17 Mind-Blowing Treasures Discovered In Thrift Stores That'll Leave You Envious
19 Heavenly Photos That Blessed My Eyeballs On This Fine Summer Day
17 Photos That Are Guaranteed To Make You Shart Your Pants In Pure Terror
Yes, shart. Journalism at its finest. The author is a “Senior Writer” at Buzzfeed.
So, let’s all please sit down and ponder what she and Buzzfeed have to say about being “wrongly accused” of “cheating.”
The piece jumps off discussing how “AI-cheating detectors aren't totally reliable” and how this leads to “false positives” and students being “falsely accused.” But then it says it has 12 examples of that:
or something really similar
Oh. So not that, but similar to that. Got it.
We begin with what should be obvious to everyone, except apparently to those who write for Buzzfeed - a student saying they did not cheat does not mean the accusation was false. Moreover, Buzzfeed is so confident that the allegations in these examples are false that the browser tab description of this article reads:
12 Innocents Students Falsely Accused Of Using AI To Cheat
Yes, “innocents.” And, as you’ll see, only one of these 12 examples is about using AI to cheat at all.
In example one - sourced from Reddit, by the way - the professor supposedly informs a student, in part:
I am emailing you because I have determined that your response … was written by AI. I’ve made this determination on my own as well as having used the GPTZero software
As you may know, GPTZero was a very bad detector. Still, I trust teachers. And if a professor, on their own, thinks something is off, it probably is. The teacher’s expert analysis and the confirmation, even by GPTZero, are strong indicators.
I won’t get into the student’s reply as it’s too long and riddled with errors. The student tries to cite the accuracy disclaimers conveniently made by GTPZero and how “two sentences supposedly written by AI” should not result in a zero on the assignment. And how they’re angry that they have to spend more time on this given their “busy schedule” and “after paying so much to attend these classes.”
Example two is from 2019, so, not related to AI detectors. But it’s a student - sourced from Reddit - who says she was accused of plagiarism by two professors and “given a big, long talk.” After showing the student the art the instructors thought she plagiarized, the student says:
it was my own art from my own site that popped up in a Google image search
Which is, in every school in America, plagiarism - even if it’s yours.
Number three, again from Reddit, also has nothing to do with AI or AI detection. In it, a professor notes that the student left an exam window many times during an exam - 13 times over three questions. The professor docked the student 20 points. The student says they did not leave the exam window.
That’s the whole thing.
Number four is from 2020 and also from Reddit. In this example, a professor noted suspiciously high scores on exams during the transition to remote teaching and assessment, and decided to curve everyone’s grade down to the program mean. The professor said he is encouraged to curve grades up to that standard if grades are low, so he feels it’s appropriate to round down in this case. And the professor does say the abnormally high grades likely indicate cheating.
The professor is probably right about the cheating. But importantly, he accuses no one specifically of cheating. The curving is applied to the whole class. But to Buzzfeed, that’s a false accusation, I guess.
Example five, also from Reddit, does not even reference cheating. The professor mentions “events during exam 4.” That’s likely cheating, but we don’t know. As a result of those events, the professor tells the class that they are dropping the results of the highest test grade and the lowest test grade. I confess I don’t see what the issue is here.
Example six - also Reddit, also nothing to do with AI. In this one, an anatomy student says they guessed the term “thyroid notch” on an exam even though the term was not in the course materials, but it was the first answer under a Google search, and was also given by several other test-takers. Quite a coincidence if you ask me.
Seems the professor thought so too and cautioned the class about cheating. Other than a strongly worded “don’t cheat” message to the entire class, it’s not clear any penalty was imposed, except maybe getting the answer wrong.
Number seven is from Twitter but the tweet and account are gone. Literally, all the Buzzfeed story says of it is:
Good morning. My instructor accused me of plagiarism, so I sent her this. Yes, I am THAT petty.
I found the “this.” It’s a message to her teacher in which she says she did not plagiarize. Pretty hostile. But that’s it - teacher said it’s plagiarism, student says it’s not.
Number eight, also Twitter, also nothing to do with AI or AI detection. It is an expletive-filled tweet, also from 2020. The short version is the professor said a student’s answers were “very close” to those found online. And that:
it was explained to you in class last week that you could not use any of the online notes. You need to redo this assignment in your own words.
The student's response is not shareable here but were I this professor, I’d remove them from class for the reply alone. I’m sure “dank ass professor” is not appropriate. And that’s not even half of it.
Still, aside from the student’s denial, I am not sure how this qualifies as a “false accusation.” And the penalty, it seems, was to do it again.
Nine - also Twitter, also nothing to do with AI, also vanished. All it says is that a student says they had been accused of using a site called StudyDaddy. It’s not clear, based on what’s there, that the student even denies it, only that they had been accused. But I am guessing that, if the professor cited the exact source, there’s probably a reason.
Ten is — and stop me if you already heard this — from Twitter and not related to AI. In this doozy, the student even says they were not even directly accused of plagiarism, tweeting:
raised concern without directly saying she thinks I plagiarized
So not only is this not necessarily a false accusation, but it’s not even really an accusation at all. And it’s not clear any penalty was imposed. Seriously, according to what we know, here is what the teacher said:
The horror.
Number 11 - Twitter, not AI - is simply someone posting that their teacher thinks they plagiarized a poem about spiders. I am mystified.
And finally, number 12 is about AI but has nothing to do with detectors or academic work. It’s a Tweet, and the exchange seems to be that a teacher wrote to a student asking that they not use AI for their e-mail messages, saying they “lack warmth” and requesting contact by a human going forward. The student replies that they are “just Autistic.”
OK. Sure. Plenty of water to tread on both sides of that exchange, but a false accusation of cheating it is not — by any measure.
After all that, we get two of 12 examples having anything at all to do with AI, and only one of those is about cheating. And the one that was related to detection was also with teacher determination plus AI detection. Every single one sourced from Twitter or Reddit and not a single one with any evidence that the accusation was false.
I know - no one wanted a dissection of some stupid article on BuzzFeed. But parts of it are funny.
And sad. That for all that unadulterated blithering garbage, we got another headline-driven, widely circulated article about the fallibility of AI detection and false cheating allegations. That matters.
A Fully-Funded PhD in Academic Integrity, In Spain
There is what seems to be a pretty exciting opportunity to study academic integrity with the University of the Balearic Islands in Spain. The application deadline is October 16.
Here is a LinkedIn post with the details, courtesy of the European Network for Academic Integrity.
Portland and the Possible Non-Zero
Let me start by saying I could be wrong about this. Hence the “possible” in the headline.
I have contacted the media relations and public information people at Portland schools seven separate times about this question. They have responded, but been unresponsive - never really answering even direct yes-or-no questions. In my latest outreach, I said I planned to write the following two sentences and asked them to correct me if I was wrong.
Here they are:
In Portland, students who are found to have cheated may be subject to discipline, but they will not receive a score of zero for the assignment. A new grading policy will avoid giving students zero credit for their assignments, even when cheating is involved.
The district has not corrected me. If they do, I will clarify and correct as needed.
The new policy aims to set the lowest possible score for any assignment to 50%. Assignments that are not completed or not submitted, would get 50% credit. So too, it seems, when cheating is discovered.
Behind the new direction is an effort to account for and calibrate student readiness and access to resources. Not all students have the environmental conditions or support needed to complete school work, and mathematically, a zero is a harsh penalty. I get it.
I think such a policy is a grand mistake though, as it’s divorced from actually living. I wish, for example, I got to keep half my house if I didn’t pay my mortgage. Or that I still get half my salary if I do not show up for work.
Even so, this is not a newsletter about grading and pedagogy, it’s about cheating. And I have a significant issue with a policy that reduces the risks of cheating, as this one does - assuming I am right.
For quite some time I have favorably cited the professor who, when he finds cheating, assesses a negative score - a minus-100 on a 100-point assignment, for example. He says that if the penalty for cheating is just a zero, cheating becomes a rational option. By that logic, if the penalty of cheating is still half credit, cheating moves from a rational option to the rational option.
Much of the recent literature around academic misconduct has isolated it as a rational choice, a clear measurement of risk and reward. Lowering the risk, as it appears Portland is doing, is frankly indefensible.
I really hope I’m just reading the reporting wrong and that someone who knows better will correct me.