What a Professor’s Chart Reveals About AI Cheating In College (And What Parents Can Do)
Last Saturday, my brother Joe, a computer science professor at our state’s public university, showed up for brunch with a crumpled printout tucked in the side of his puffer jacket, a ring of coffee stain blotted across the bottom corner. He’d stopped by on his way to a faculty meeting, still annoyed from his 8 a.m. midterm review, and spread the paper across my kitchen table between the plate of bacon and the pitcher of iced tea. It was a simple grid: rows for each of his 280 intro to coding students, columns for every assignment this semester, cells shaded pink for confirmed AI use, yellow for questionable.
The midterm column was half pink.
Forty-one percent, Joe said, tapping the grid with his coffee mug. Three years ago, before the current mainstream tools, that number would’ve been under 3 percent. Two years ago, it was 12. Last year, 27. Now this. He’s been tracking it every semester to present to the curriculum committee, trying to get the school to acknowledge how big the problem actually is—most professors just guess, or don’t report cases because the disciplinary process is a nightmare for everyone involved.
My daughter Lila, a junior at the same university, was leaning against the counter eating a slice of toast, and she chimed in before I could ask anything. It’s not just comp sci, she said. In my human development midterm, the take-home essay? I’d bet at least half the class used AI to write it. I asked her why. It’s not like most of these kids can’t write their own essays, right? She shrugged, wiped crumbs off her jeans. Everyone else is doing it. If I spend 10 hours writing mine and someone else uses AI for 20 minutes and gets a better grade, that hurts my GPA, which hurts my shot at grad school. Tuition’s $15,000 a year for in-state, more for out. No one can afford to get a B when an A is a few taps away.
That’s the part no one talks about enough, Joe said. It’s not that this generation is lazier or less honest. The incentives are all messed up. We have 280 kids in an intro class that’s a weed-out for the major. We grade on a curve. The kids know that even if they know the material, if everyone else is cheating, their grade is going to be lower. So they cheat too, just to keep up.
As a high school teacher, I get this question every week from parents: is my kid cheating with AI? What do I do about it? Most of the time, the first thing parents want to do is confront them, accuse them, or put rules in place that don’t work once your kid is living on a college campus three hours away. After brunch, when we were walking Joe to the train station, we talked about what actually works, and it’s nothing fancy.
I’ve started doing this with my own students and with Lila, and it’s way more useful than yelling or demanding answers. Next time you’re talking to your kid about school, instead of asking “did you use AI to cheat on that assignment?” ask them: “Walk me through what you use AI for when you’re stuck on work.” It’s an open question, not an accusation, and it gets them talking instead of shutting down. Last month, Lila told me she uses AI to debug her code when she’s been staring at the same error for an hour. She’ll paste the broken code in, ask it what’s wrong, then fix it herself. That’s not cheating—that’s using it like a study tool, same as we used Google or a tutor when I was in college. The problem is when kids use it to write the entire answer or code from scratch, and that distinction almost always comes out if you ask how they use it, not if they use it.
The other concrete thing you can do is shift the focus away from just grades, in small, daily conversations. Instead of asking “what grade did you get on your midterm?” ask “what’s one thing you learned this week that you didn’t know last month?” It sounds small, but it changes how your kid thinks about what school is for. Joe told me the kids who cheat on intro coding exams almost always flunk out of the upper-level classes that build on that intro knowledge. They didn’t learn the basics, so they can’t keep up when the work gets harder, even if they cheat again. If kids are focused on learning the skill instead of just getting a good grade, they’re way less likely to cut that corner, because they know they’re only hurting themselves later.
For high school teachers and college professors who are reading this, Joe’s small adjustment to his exams has cut AI cheating by about a third, and it’s not hard to implement. He now has half the exam grade come from a 45-minute live coding section done on a department computer that doesn’t have internet access, only the class documentation installed. No phones allowed at the stations, you have to write the code from scratch in front of a TA. It doesn’t stop all cheating, but it makes it way harder to sneak AI in, and it tests whether you actually know how to write the code, not just whether you can copy it down from your phone.
Joe folded his crumpled chart back into his jacket before he got on the train, and said he doesn’t expect the university to make big changes any time soon. They don’t want to fail 40 percent of an intro class, they don’t want the bad press, they don’t want to deal with the pushback from parents and students. It’s easier to just pretend the number is lower than it is.
I’ve been thinking about that chart all week. I have 32 seniors in my classes this year, most of them are heading to college next fall. I’ve started adding that five-minute check-in with each of them before they leave for graduation, talking about how to use tools honestly, how to not get pulled into the race to cheat just because everyone else is doing it. It’s not a perfect solution. There will still be kids who cheat, there will still be professors who don’t want to talk about how big the problem is. But it’s a start, one conversation at a time.
Please indicate: Thinking In Educating » What a Professor’s Chart Reveals About AI Cheating In College (And What Parents Can Do)