Chegg CEO Says Schools are "lazy," and He Doesn't "give a sh*t about” What Kids at Harvard Did
Plus, an update on the UNC Professor and his course to make students not want to cheat.
Issue 224
To join the 3,373 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 a month via Patreon. Or joining the 14 outstanding citizens who are now paid subscribers. Thank you!
Fortune Magazine: Chegg CEO Spars With Gen-Z Former Student Over ChatGPT, Learning and More
Fortune Magazine has a story out today (subscription required) about an event this week - Fortune’s Brainstorm Tech conference, which was apparently attended by Chegg’s CEO Dan Rosenweig. The magazine said he was there to promote Chegg’s new AI product, CheggMate (see Issue 203).
There, Rosenweig was seemingly confronted by a former Harvard student and self-identified “Gen Z elder.” The two, according to the reporting, clashed over education, learning and ChatGPT/AI. The magazine framed it as “Generational attitudes about A.I. in the classroom were turned on their head.”
I’m not sure that’s true. I think young people are in on the joke - that OpenAI and Chegg and other easy answer providers are not learning tools, that they are cheating shortcuts. I’ve said this before. It’s the old people who make or sell those products who tell themselves and their investors and the media that just giving answers is some new kind of way to learn.
But none of that really matters here because, in the minimally reported dialogue of the exchange, it’s clear that a provoked and passionate Rosenweig said some things. Those are interesting.
To set it up, the 2021 Harvard grad and non-profit founder said, apparently to Rosenweig:
I was a Chegg user but not because I gave a f*ck about learning
I added the asterisk. This is a family newsletter.
She continued that she was a Chegg user:
because it gave me answers to problem sets. I meet a lot of young students out there that aren’t necessarily interested in using ChatGPT for learning. They’re using it because it makes it easier to complete homework.
Yes indeed. That is why people use Chegg. And ChatGPT. They are not interested in learning with Chegg. Say it with me - Chegg is not a learning tool. It is not an education company. It is a ‘make it easier to do your homework’ company. Or tests. Whatever.
Fortune said the former student, Nadya Okamoto:
was skeptical of Rosensweig’s hopes that Cheggmate or any other A.I. tools would be used exclusively in the pursuit of knowledge. These tools, she said, ultimately create little incentive for students to actively engage with their course materials.
No kidding. Seems obvious - why learn when Chegg or ChatGPT will just spit the answers to you?
She continued:
A concern of mine, as a Gen Z elder, is feeling like these sorts of tools are making it easier and easier to honestly just get the answers and deprioritize engaging curiosity
Ding, ding, ding.
She went on to say that she used Chegg in her junior year because it was “easy” and because, “C’s get degrees.”
Fortune reports that Rosenweig said in response that Okamoto was:
thinking about it from the exact wrong way
The magazine characterized his response further as:
explaining that Chegg was created for the type of student who doesn’t have the same level of resources a student at an elite university does.
Which is odd, since Chegg sells subscriptions to Harvard students. And they use it. But if they don’t need it because they have good resources, I am baffled as to what these students are actually buying. Hmm. A genuine mystery.
Quoting Rosenweig further:
“So what the kids at Harvard did, I don’t give a sh*t about,” he said. “I care about the kids who are trying to make their life better.”
Again, my asterisk. But Harvard “kids” are not trying to make their life better? Which, again, if that’s the case, why sell Chegg to Harvard students?
It’s rhetorical. We all know why.
But do please note that a former Harvard student told Rosenweig, to his face, that she used Chegg to get easy answers to her assignments. Rosenweig’s response: I don’t care.
Rosenweig also repeated the invented line that Chegg was for non-traditional students who were “unsupported by [their] school system” and maybe had children or demanding jobs. Those, according to Rosenweig, are the people who should pay $19.95 a month for Chegg.
Rosenweig continued, from Fortune:
“My view is we are built for the student that is self-initiated,” he said. “That wants to learn, and that needs to graduate with a skill.”
He went on to call academic institutions “lazy,” chastizing them for not updating their curricula and professors for not writing more thoughtful exam questions that couldn’t be answered by a ChatGPT prompt.
This makes me laugh. Like, out loud.
To start, non-traditional students who are self-motivated and want to learn do not need to pay for a Chegg subscription. A library card is free. So is YouTube. And Google. So are the academic resources at their schools. Or at least they’re included with the tuition.
If you are “self-initiated” you not only don’t need to pay Chegg, you don’t want to pay for Chegg. It is not a place you go to learn. As the former student quoted in this piece said, it’s a place you go for easy shortcuts.
Which makes it ironic that Chegg’s CEO calls schools lazy. Catering to lazy is his business model.
And finally, I am certain that Chegg’s CEO would love it if teachers started writing exam questions that can’t be answered by ChatGPT. He’d love it if students could not get their answers from ChatGPT because ChatGPT is killing Chegg (see Issue 206).
Follow-Up: The UNC Professor Who Said “Did Some Students Cheat? I Hope Not.”
In Issue 220, we shared the story of a professor at University of North Carolina, Chapel Hill who wrote an article in praise of his teaching and course design that he said would make students not want to cheat.
He wrote:
But rather than spending a lot of time trying design courses and exams where students cannot cheat using ChatGPT, I am spending my time developing courses where they do not want to.
As reminder, the course he co-taught had 333 first-year students. I was skeptical. I still am.
So, I asked the University for more information about the assessments in the special ‘students don’t want to cheat’ class. They kindly sent me the syllabus and some comments from the professor.
According to the syllabus, 30% of the course grade was based on ten journal entries of between 300 and 500 words. One a week. Points were deducted if an entry did not have a “clear title.” It is not clear whether these were assessed for possible plagiarism or AI-creation.
Then, 20% of the final grade was the midterm, about which the syllabus says:
The midterm exam will be a combination of multiple choice and short-answer items. The questions will be derived from assigned readings and material covered in class. The midterm will be taken at home, via Canvas, and must be submitted by March 9th, 2023 at 5:00pm EDT. The midterm exam will be worth 20 points. Students are expected to complete the exam in 75 minutes without any resources or aids (i.e., it is “closed book”).
I asked a few times whether this exam was proctored or supervised. I never got a real answer so, no. It was not.
Which means that the professor who was proud of designing this course gave students an online, unproctored, open-ended, at-home exam with multiple choice and short answer, which he “expected” to be done “without any resources or aids”.
Sure.
When pressed a bit, a University spokesperson said:
The midterm was online, and students completed it on their own. They were asked to confirm adherence to the UNC-Chapel Hill Honor Code, which governs the performance of all academic work a student conducts at the University. All students on our campus agree to adhere to the honor code when they attend the University. There are consequences for those who do not adhere and are found to have cheated.
Problem is, of course, you cannot know if anyone cheated if you’re not paying any attention. Giving 333 freshmen an unsupervised, on-line, asynchronous midterm exam - I mean, if you wanted students to cheat, that’s exactly how to do it.
The final exam was worth 30% of the grade and was “a format similar to the midterm.” Though the professor, in response to my queries, did say the final was proctored, in-person, which highlights that the midterm and other assignments were not.
Further, as part of his response to my query about his class, the Professor wrote:
we did our best to create a course and classroom environment where students understood that they were impeding their expertise development when they yielded to temptations like ChatGPT, or any other of the ways students have cheated in the past.
There is exactly zero chance that worked.
Moreover, he and his co-teachers did not do their best to prevent or dissuade cheating. Not at all. Not even remotely close. On the contrary, it’s pretty clear he doesn’t care about it.
He essentially said so. In his article for his local paper, the professor wrote:
But I am not worrying about students cheating with tools like ChatGPT. Instead, I am focusing on making classes where students do not want or need to cheat, at all.
Professors who don’t worry about students cheating, put no effort into addressing it, and think they can create a course ‘where students understand that cheating impedes their development’ - it’s a problem. And I hate blaming teachers for student cheating. But on occasion, that is exactly where it belongs.