"If the Public Lose Faith in What We Do, We Lose"
Plus, another flyer advertising cheating services arrives in Oklahoma. Plus, ICAI calls for award nominees.
Issue 258
To join the 3,697 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 27 outstanding citizens who are now paid subscribers. Thank you!
We Lose.
The quote in the headline — if the public lose faith in what we do, we lose — is from this article, from the Australian Broadcast Corporation. For the most part, it’s really strong, and important.
Jumping off on the accepted idea that teachers and schools braced for a wave of cheating after ChatGPT emerged, ABC says the wave did not really happen. Though, they crucially add:
Many teachers said they strongly suspected students were cheating, but it was hard to tell for sure.
Based on what we know from multiple survey data points, the teachers were right, students were cheating. It just was not being caught. They still are. And for the most part, they still aren’t.
Doing some actual reporting, ABC talked with high school students about cheating and GPT. They write:
Some were initially curious, but also cautious.
Some used it once then stopped. Others kept going.
Some got caught, but many didn't.
Exactly.
Recounting an interview with one student, ABC reports:
In term one, he experimented with cheating, using ChatGPT to write a take-home geography assignment (though it didn't count towards his HSC).
He wasn't caught and got a decent grade.
Next, he gave ChatGPT his homework.
"It's just meaningless, monotonous work. And, you know, the chatbot could just smash it out in seconds."
Eric estimates that, over the course of 2023, ChatGPT wrote most of his homework.
"So around 60 per cent of my homework was written by ChatGPT," he says.
Yup. He used it, did not get caught, used it more. He decided on his own that the schoolwork was “meaningless” and let AI do it, learning nothing, getting decent grades while doing no actual work.
Eric says also that:
the AI tool provided a shortcut through an already flawed system.
ABC says:
Students at other schools told similar stories.
No kidding.
They go on:
Chrysoula, a Year 11 student in Sydney, initially used ChatGPT "very often to complete homework I deemed just tedious".
A lot of her classmates were doing the same, she says.
Again - she deemed it “tedious,” so she did not do it. Classic rationalization. To which, ABC adds:
Student accounts suggest that, although not all want to cheat when they can, many are happy to automate tasks they see as rote learning rather than true tests of knowledge.
I am sure that every teacher on Earth is just fine with students deciding on their own which assignments are rote or tedious or meaningless and simply not doing them. What could be the trouble with that?
On a separate point, ABC reports:
Many of the students we spoke to say teachers have little power to stop them from using AI tools for homework or assignments.
And:
Students described numerous ways of evading the blocks on accessing ChatGPT on school computers, or through a school's Wi-Fi network.
They also told how they copied their ChatGPT-written answers into other AI tools, designed to confuse the schools' AI-detection software.
Yes, that’s what the Course Hero owned Quillbot does. And others. They exist to hide plagiarism and AI text. Students know it. They use it.
Also, those in academia who worry about creating and keeping positive learning cultures and healthy in-class dynamics should perk up at this, from the same student, Chrysoula. Saying that using AI was so common, she says:
"Everyone doubted the authenticity of everyone's answers."
Tell me how that’s healthy.
There’s more:
Phil, a Year 12 student at a different high school, also sees a divide that's formed among his classmates, based on how much each student uses AI to do their coursework.
His school allows students to use ChatGPT for ideas and research but not to directly write assessments.
But many students still cheat on take-home assignments, Phil says.
I’ll say it again - you know who hates cheating more than teachers? Students who don’t cheat. Failure to protect honest work is the greatest injustice of academic integrity inaction.
Later in the article, ABC quotes the author of a paper on AI use in school settings, who says:
teachers and parents consistently overestimate how many students use AI to cheat.
"While 58 per cent of students report having used generative AI in the past year, only 19 per cent of that group reports using it to write and submit papers," she says.
We know that self-reported misconduct under counts it, so 19% is probably on the low end. Even so, let’s pretend that number is accurate. It’s still 11%. More than one in every ten students admits to using AI “to write and submit papers.”
And we’ve all seen other student survey data which puts that number comfortably at or even north of 30% (see Issue 190).
But allow me to ask a question. Is the bigger issue that teachers and parents overrate cheating with AI, or that 11% admit to doing it? Assuming it’s only 11%.
Anyway, ABC gets more credit for interviewing an actually credible source on academic integrity - Kane Murdoch of Macquarie University. He told ABC that even if the cheating rates are low, they are likely to go up:
He believes students will gain confidence, and learn how to use AI to automate more of their coursework.
"It could be 2023 was the year they dipped their toe in the water, but 2024 and moving ahead you’ll get increasingly large body parts into the pool.
"And soon enough they’ll be in the deep end."
Murdoch errs somewhat when he adds his voice to the chorus of doubt regarding the ability to detect AI in student work. He says:
"There's lot of skepticism about the efficacy of detection — and I'm among those who are skeptical."
He says educators were reluctant to rely on flawed plagiarism-detection tools to accuse a student of cheating.
"As an investigator [of cheating] I'm unwilling to accept the detectors word on it," he says.
All I can add is that the evidence is pretty clear (see Issue 253 or Issue 250). And that he’s right — educators should be reluctant to rely on AI detection. They should never use it by itself. No educator should use AI detection as the only evidence. They are detectors, not deciders. They should start conversations, not end them. I really thought we were past this.
ABC also reports:
ChatGPT-maker OpenAI has warned that there is no reliable way for educators to work out if students are using ChatGPT.
OpenAI did say that. But they are wrong (see Issue 241).
And the clincher, as far as I am concerned, is from Murdoch, who tells ABC:
He's concerned that AI cheating will ultimately devalue educational qualifications.
"We can't expect this to go away. It is a game changer — it is existential," he says.
"If the public lose faith in what we do, we lose."
He’s right.
AI cheating — all cheating, I’d say — devalues education qualifications. It is existential. If we don’t start taking it seriously, we lose. And make no mistake, we are losing.
More on Cheaters Advertising, By Mail
In Issue 248, we covered the story of someone who received an ad from an essay mill selling cheating services — in the mail.
I speculated at the time that though that incident took place in Utah, where cheating is legal, it was likely that the same or similar ads were mailed to other places.
Well, according to this news story, a similar flyer for cheating services landed in the mailbox of a retired professor in Oklahoma. According to the reporting:
“Why are you still doing your own work?” the flyer asks.
And:
The website’s owner doesn’t hide the fact that the provided services help students cheat on their work.
“The longest paper wrote was 45 pages,” one TikTok post said. “But my client is still in school, so I don’t want to post that yet because I don’t want to get anyone in trouble. Like, everything is 100 percent confidential because college is a scam, so I’m gonna help ya’ll cheat ya’ll way through it.”
Typical.
What’s more noteworthy to me is that this retired professor:
checked with the United States Postal Service, and the information he received was that the flyer was most likely a fraudulent piece of mail.
“I first went to the post office, and they got the manager,” McCoy said. “His quote was, ‘This is a crime, this is criminal stuff.’ It’s stealing information and lying about where you got it.”
As mentioned in the previous coverage from Utah, using the postal service to sell cheating could have actual consequences. Though it’s pretty obvious that, apart from our noble former educator, Jerry McCoy, no one really cares.
That noble educator, meanwhile, also told the local paper:
“This is the devastation of the truth of your value of your undergraduate and graduate work,” McCoy said.
And:
By not legitimately completing courses, students are lying to themselves, he said.
“You’re not who you say you are,” McCoy said. “This is an undoing of the whole concept of truth in education.”
He’s right, obviously.
Though I am sure I am also right. No one cares. At least not enough to do anything.
ICAI Calls for Award Nominees
The ICIA, International Center for Academic Integrity, has called for nominations for its annual awards.
The link to review and submit nominees is here: https://mailchi.mp/academicintegrity.org/we-need-your-assistance-2618400?
The ICAI Annual Conference is in March, in Alberta, Canada.