The Brown University Cheating Scandal

There's been a story out over the past week or so about a cheating scandal at Brown University. It's behind a paywall at the incongruously named Free Press, but the gist of it is at USA Today.

There's a new morality emerging in the corridors of America's finest universities: If you're not cheating, you're a sucker.

Look no further than Brown University, where an artificial intelligence cheating scandal has rattled longtime economics professor Roberto Serrano. This spring, after three decades of teaching at Brown, Serrano let his students take their midterm exam at home, upping the difficulty to compensate for the unlimited time. The average score, which had historically run between 65% and 80%, jumped to 96%.

Was that a particularly studious class? Maybe. But Serrano suspected a simpler explanation: AI.

He announced that the final would be given in person, causing 18 students to drop the class. Nine more didn't bother showing up for the exam.

The problem I see here is that if people are relying on AI to get ahead, they're going to be replaced by AI. This reminds me of a line from anti-woke comedienne Arynne Wexler, who says that her job -- anti-woke comic -- won't be replaced by AI, because AI isn't racist. The USA Today piece continues,

A recent survey of more than 95,000 undergraduates at 20 American public research universities estimated that 9% used generative AI to cheat.

Among daily AI users, that rate jumped to 26%.

Amy Clukey, an associate professor at the University of Louisville, estimated that more than half her students in one online course had committed AI-facilitated plagiarism.

I don't know where people are getting these numbers. When I was teaching freshman comp as a graduate assistant at USC 50 years ago, the consensus among all of us was that plagiarism was much closer to 100%, and this was without AI. Cheating was nothing new. Although it's not online, I clearly remember an op-ed by Gerald Seib at the Wall Street Journal in the 1970s asking, "Why Are They Cheating?" giving a very similar estimate. In other words, this is much ado about nothing new.

And Prof Serrano just gives a new example of the old economic incentive for professors to tolerate cheating: any serious effort to control it will result in a drop in class enrollment. The USC English Department was fully aware of this 50 years ago and simply turned a deaf ear to the graduate assistants' complaints. Nothing has changed here; if Prof Serrano continues his efforts, his department chair will notice the enrollment drop, and things won't go well for him. USA Today claims this is a new development:

Amanda Gregg, an associate professor at Middlebury College, told The Chronicle that her students "acknowledge that there is not a culture of social sanctions against cheating." Gregg recalled that when she was a student, cheaters faced “major social sanctions.” When she mentioned this to her own students, she said, they were “shocked.”

. . . We're witnessing a move away from a culture that valued honor and integrity (think of the social conditions that once made “a man of his word” mean something) toward one where ambition and prestige are the only markers of worth. Call it “striver culture.” To a striver, the ends always justify the means.

To be a striver is to be concerned with the appearance of prestige above all else. It means attending the high-status school, getting the high-status job, being friends with high-status people.

Sinclair Lewis's George F Babbitt advised his son in the 1920s:

"I'll tell you why you have to study Shakespeare and those. It's because they're required for college entrance, and that's all there is to it! Personally, I don't see myself why they stuck 'em into an up–to–date high–school system like we have in this state. Be a good deal better if you took Business English, and learned how to write an ad, or letters that would pull. But there it is, and there's no tall, argument, or discussion about it! Trouble with you, Ted, is you always want to do something different! If you're going to law–school—and you are!—I never had a chance to, but I'll see that you do—why, you'll want to lay in all the English and Latin you can get."

Oddly, I just don't see the most effective line that could be given by professors against using AI to cheat: if you have to use AI to get by, you're going to be replaced by AI. But if a professor actually designed a course that incentivized not using AI, he would, like Prof Serrano, see an enrollment drop. Sinclair Lewis, who went to Yale, seems to have had a pretty good grasp of what college has always been for.

As I've noted here, I've restarted a fiction writing career, and what I'm noticing is that all the little lit zines have severe anti-AI policies -- they don't accept submissions that have been written by AI, except that all AI really does is take the formulas that are taught in creative writing class and follow them. The little lit zines by and large want stories written the way they teach in creative writing class. Just last week, I had to deal with an editor who wanted to adjust my story to better match a creative writing prof's idea of "story arc".

All AI is doing in this case is giving people what they want, but doing it more cheaply. In college, students are using AI to give the professors what they want, but doing it more cheaply. What's wrong with that?