How Common Is AI Cheating In College Exams? What A Professor’s Data Teaches Parents
I was wiping blueberry muffin crumbs off our kitchen table last Sunday when my daughter Lila, home for spring break from her junior year of college, tapped her phone and slid it across the table to me. “Look,” she said. “This is my stats prof. He made this chart showing how much AI cheating is actually happening right now.”
The x-axis ran from 2018 to early 2024, the y-axis the percentage of exam submissions flagged by AI detection tools. The line was almost perfectly flat, hovering right around 1%, from 2018 through the end of 2022. Then it turned straight up: 6% in spring 2023, 12% that fall, and 19% of all exam submissions last semester. That’s almost one in five. The chart didn’t leave any room to downplay how big the shift has been, no matter how much schools have tried to ignore it.
Lila leaned back in her chair and shook her head. “Everyone on campus is talking about it, but it’s not what the headlines say. It’s not all just lazy kids trying to get an A without working. My roommate got flagged last semester because she used ChatGPT to outline her midterm essay after pulling an all-nighter with the flu, and she forgot to rewrite the body paragraphs in her own words. She got put on academic probation, and she wasn’t even trying to cheat.”
As a 12th grade English teacher, I’ve been watching this trend trickle down to my classroom for two years. Kids get comfortable using AI for homework and study help in high school, then show up to college with no clear idea where their school draws the line. Before I saw this chart, I thought the problem was overblown, that it was still just a small group of kids cutting corners. This data changed my mind: it’s not a handful of bad actors, it’s a widespread shift in how students work, and most schools haven’t caught up.
I used to just tell my own kids and my students “don’t use AI to cheat” and leave it at that, but that’s useless advice right now. The line between allowed and not allowed isn’t clear anymore. Some professors let you use AI for brainstorming, others ban it entirely. Some high schools allow it for editing, others treat any use as cheating. Now I do two small, concrete things that have cut down on confusion a lot. First, on the first day of each semester, I walk my students through highlighting the AI policy in the syllabus together. Most kids skip all the fine print when they get a new syllabus, so 10 minutes of class time to pull out the rule and talk about what it actually means clears up 80% of the accidental missteps. As a parent, I started doing the same with Lila: we sit down for 5 minutes at the start of each semester and go through each of her course syllabi to mark the AI rule together. It takes no time, and it keeps everyone on the same page.
For my own tests and essays, I stopped trying to ban AI entirely and changed what I ask students to do. Instead of asking them to write an essay arguing for or against a common topic, I ask them to walk me through how they changed their mind while they worked, or explain what part of the research they found most confusing. AI can spit out a coherent argument, but it can’t explain how I got stuck on the second paragraph and went back to re-read the source to change my thesis. Last year, after I started doing that, the number of AI-flagged submissions in my classes dropped by 70%. Kids still use AI for brainstorming, but they have to do the hard, personal thinking themselves to get a good grade.
The professor who made that chart didn’t release it to call out students, actually. He wrote in his accompanying post that he’d reviewed almost all the flagged submissions, and only 3 out of 10 were premeditated cheating. The rest were accidental, or kids who misread the rules, or kids who were so overwhelmed with midterms and part-time work and family stuff that they cut a corner they never would have otherwise. That matches what I see too. Most of the kids I see cross the line aren’t bad kids, they’re tired kids who don’t have anyone telling them clearly what’s okay and what’s not.
I get why parents panic when they hear about 1 in 5 college exams having AI cheating. I’ve had parents come to me after parent-teacher conferences, asking if their kid’s AP essay grade even means anything anymore, if they’re actually learning anything. The answer isn’t to panic and ban AI entirely, or to pretend the problem doesn’t exist. It’s to stop treating this like a moral failure and start treating it like a learning curve, for kids and schools alike.
Before Lila left to drive back to campus that Sunday, we stuffed the leftover muffins into a Tupperware for her to take. She said her stats prof is using his chart to push the college to update its academic integrity policies and offer more workshops for students on how to use AI properly, instead of just increasing punishments. That makes sense to me. I still have days where I worry that we’re fumbling through this, that we’re not getting it right. Last month, one of my top students got flagged for AI on a take-home essay because he used AI to fix his grammar, and his teacher had a zero tolerance policy. We spent two hours meeting with the principal to get it sorted out, and that wouldn’t have happened if the rules had been clear from the start. Things are changing faster than most schools can keep up, and there’s no neat fix right now. But that chart made one thing clear: pretending the problem is small isn’t working anymore. We have to start talking about it, clearly, with our kids, before more of them get caught on the wrong side of a line no one ever bothered to draw.
Please indicate: Thinking In Educating » How Common Is AI Cheating In College Exams