With AI in the College Classroom: Why Professors Are Walking a Tightrope (And What It Means For Students and Parents)
It’s 3:15 on a rainy Tuesday, the office coffee pot has been empty for 40 minutes, and a junior named Mia is standing in my doorway twisting the strap of her canvas tote. She’s got her first draft of her food desert research paper tucked under her arm, and she looks like she’s about to cry. She says, “I need to tell you something. I used ChatGPT to sort my USDA data and make my first outline.”
I remember the department memo we got last fall: any unreported AI use results in an automatic failing grade for the course, referred to the academic integrity board. Disclosed use is still supposed to get a minimum zero on the assignment. But Mia already told me at the start of the semester she works 25 hours a week at a daycare off campus to cover tuition, that she’s supporting her little sister who stays with her on weekends. That week, she’d missed earlier office hours because she had two midterms back to back.
I could send her to the board, fail her, and check the box for academic integrity. Or I could cut her a break, and worry that I’m letting her skip the hardest part of the assignment—learning how to organize messy research into a coherent argument. That’s the tightrope we’re all walking right now. One wrong step, and either a kid’s future gets derailed, or they leave my class without the skills they paid to learn.
Three weeks ago, we sat through a two-hour department meeting about AI policy. The new tenure-track professor hired last year said he bans AI entirely, runs every paper through three different detection tools, and fails anyone that comes up positive. But we all saw what happened last semester when one of those tools flagged a senior’s creative writing thesis as AI-generated. She’s a quiet kid who writes really straightforward, plain prose, and she had to spend three days proving she wrote every word. She almost didn’t graduate. The detection tools are wrong 30% of the time, at least, according to the studies we’ve all seen. No one wants to talk about how often they punish kids for nothing.
Another professor there, who’s been teaching 30 years, said he just makes students bring all their notes, drafts, and scratch work to class when they turn in a final. He doesn’t care if they used AI—he just wants to see how they got from the first idea to the final paper. That stuck with me.
I had Mia pull a chair next to my desk, and I asked her to walk me through her outline line by line. Why did she lead with co-op access instead of income levels? What made her cut the section on federal policy that AI suggested? She talked for 20 minutes, pointed out where she’d rearranged the whole second half, explained that she’d interviewed three co-op volunteers over the weekend that added a whole new angle AI didn’t know about. It was all her argument. AI just sorted the messy spreadsheets she’d downloaded from the USDA so she didn’t have to spend 10 hours doing that.
I knocked 5 points off her final grade for the assignment, because she didn’t practice the outlining step on her own first, and that’s a skill she needs to practice. I didn’t fail her, I didn’t send her to the board. That was my call, and it’s one I lose sleep over sometimes.
As a parent of a high school senior getting ready to apply to college next fall, I get why parents worry about this. I’ve had other parents ask me if professors are just letting kids cheat now, if their four-year degree is going to be worth nothing because everyone uses AI. The truth is, it’s not that simple. Banning AI doesn’t work—kids will just use it and not tell you. And zero tolerance policies hit low-income kids the hardest, the ones who can’t afford to spend 10 hours sorting spreadsheets when they’re working 25 hours a week to pay rent.
What’s the practical stuff that works, for both professors and parents? For professors, I’ve started building small process checks into every paper assignment. 10% of the grade for a 1-paragraph topic proposal due two weeks before the draft. Another 10% for an outline with source notes. That way, I see their thinking long before they turn in a final paper, and they can’t just turn in something AI wrote from start to finish without me catching it. It’s a little more work, but it’s way more reliable than any detection tool.
For parents of college students, instead of asking your kid if they cheated with AI, ask them to walk you through the main point of their last paper, and tell you how they came up with that point. If they can explain it to you, they learned something—AI just did the boring busywork that doesn’t matter that much anyway. If they can’t explain it, that’s when you can encourage them to talk to their professor about going back through the work to understand it better.
Last week, Mia turned in her final paper. It was one of the strongest I graded all semester, full of specific details and original insight she got from her interviews. She told me she’s going to use it as a writing sample when she applies for public health internships this summer. I still don’t know if I made the right call with that 5-point deduction instead of a zero. Last month, I failed a freshman who turned in a whole introduction written by AI and didn’t disclose it, and he sent me a long angry email saying I was out of touch.
There’s no official rule book for this right now. Every professor is making their own calls, every day, balancing the need to keep academic standards high with the need to not ruin a kid’s semester over a tool that’s not going anywhere. Right now, asking students to show their work instead of just banning the tool is the best I’ve got. It’s not perfect. But it’s better than falling off either side of the rope.
Please indicate: Thinking In Educating » With AI in the College Classroom: Why Professors Are Walking a Tightrope (And What It Means For Students and Parents)