Why Professors Are Walking a Tightrope With AI In the Classroom
Last Tuesday, I wrapped my 10 a.m. Romantic poetry lecture and headed back to my tiny third-floor office in the oldest building on campus—heat never works right, windows stick when you try to open them—when Lila knocked before I could even plug in my space heater. She was twisting the strap of her canvas tote, the one with a peeling campus animal rescue sticker, eyes shiny, and said she needed to ask something before she turned in her 8-page midterm.
She’d used ChatGPT to outline her paper. She’d had a bio midterm the same week, her mom had gone in for a minor outpatient procedure the weekend before, and she’d gotten so overwhelmed sorting her messy lecture notes that she’d turned to AI to organize her arguments. The syllabus said all AI use had to be disclosed, so she was telling me now, before she turned it in, and she was pretty sure she was going to fail. She just wanted me to know she didn’t copy anything—she’d rewritten every paragraph, added all her own close readings of Wordsworth we’d talked about in class, she just needed help sorting the mess.
I’m an adjunct professor here, and I’m also a mom to a 16-year-old junior in high school, so I stood there behind my desk with its chipped mug full of dead markers and felt that familiar tightrope tug: one side, I’m supposed to enforce academic integrity, crack down on plagiarism, make sure students do their own work. The other side, I know what it’s like to be an overwhelmed student with a hundred things pulling you in a dozen directions, and I’ve watched my own kid use AI to work through algebra problems when I couldn’t remember how to do factoring and we couldn’t afford a last-minute tutor. Where do I step?
Last semester, I had a student turn in a full 10-page paper that was entirely AI-generated, no original thought, no specific references to our class discussions, just generic fluff about Romanticism. I failed him, per department policy, and he came to my office hours saying everyone uses AI now, so why was he the one getting punished. I didn’t have a good answer for him. A semester before that, a junior with dyslexia told me he’d used AI to fix his grammar and spelling, because his processing issues made it impossible to catch every error after he wrote his analysis. He turned in some of the most original thinking I’d read all year, and if I’d banned AI outright, he would have either failed or had to drop the class, which didn’t seem fair either.
I’ve tried small, concrete changes instead of banning AI entirely, which is what I tell other teachers and parents when they ask what I do. First, I rewrote my syllabus to say exactly what’s allowed and what isn’t, no vague language. Allowed: using AI to organize notes, fix grammar errors, generate alternative argument ideas to explore. Not allowed: having AI write any full paragraph that you turn in as your own, or failing to disclose any AI use. That takes the guesswork out for students, instead of them having to wonder what counts as cheating. Second, I split my big end-of-term papers into three small checkpoints: first, a 200-word outline that requires two personal responses to the primary text, then a rough draft of one full body paragraph with a quote you picked and your own explanation of why it matters, then the final paper. Students can use AI to help with the first step, but they can’t get away with having AI write the whole thing because I see their work along the way.
For parents, I have a simple rule I use with my own son that I share. If your kid says they used AI for an assignment, don’t jump straight to accusing them of cheating. Ask three specific questions: What did you use it for? What parts of this did you write or think through on your own? Does your teacher allow this kind of use? That starts a conversation instead of a fight, and it helps your kid think through the line between using a tool and letting someone else do your work for you. With my son, we have an extra rule: if he uses AI to make a study guide, he has to rewrite all the key points in his own words before he studies them, so he’s actually processing the information, not just copying AI’s work.
Back to Lila: I flipped through her draft, saw she’d crossed out half the bullet points the AI generated, scribbled her own notes in the margins about how Wordsworth’s walking tours shaped his idea of nature as a teacher. I told her that as long as she added a one-sentence disclosure at the end saying she used AI to outline, the paper was fine, and I’d grade it just like any other. She exhaled so hard I thought she’d blow my stack of syllabi off the desk.
That said, it never feels perfect. Admin still sends out weekly emails reminding us to crack down on AI plagiarism, and I get it—students who cut corners get an unfair advantage over students who spend hours working on their papers. But zero-tolerance policies just push AI use underground, so students don’t disclose it when they use it responsibly, and they don’t learn how to use it well. Some semesters I fail one or two students for full AI plagiarism, and every time it feels like a failure of my assignments, not just theirs—if I’d structured the assignment to require showing work along the way, they wouldn’t have been able to turn in a fully AI paper in the first place.
I still don’t have all the answers. Some days I feel like I’m leaning too far one way, some days too far the other. Last night, my son asked if he could use AI to brainstorm topics for his college application essay. I told him he could, as long as he wrote every word of the final draft himself. We’ll see how that goes. For now, that’s all any of us can do: one step at a time, trying not to fall.
Please indicate: Thinking In Educating » Why Professors Are Walking a Tightrope With AI In the Classroom