Why Your Teen’s Assignment Keeps Getting Flagged for AI (What I Do As A Teacher Mom)
It’s 8:30 p.m. on a Tuesday, and my 10th grader, Mia, is shoving her laptop across the kitchen table at me. A half-eaten bowl of Goldfish crumbs spills off the edge onto my chipped coffee mug, and she’s got that signature 10th grade eye roll that means she’s one wrong comment away from stomping up to her room. The screen is pulled up to her To Kill a Mockingbird character analysis, and big, bold blue text across the top reads: High probability of AI generation.
“I wrote every word of this,” she says. “I even stayed up an extra hour adding that part about Scout’s shoes being a symbol. Why does this thing think I cheated?”
I get it. I’m Mia’s mom, but I’m also a high school English teacher at the nearby public high school, and I see this at least once a week. Last semester, one of my 11th grade students, Javi, got an automatic zero on his personal narrative about moving from Chicago to live with his grandma after his mom lost her job. The detector flagged it 90% AI, and the district’s new automatic grading system dropped the zero in his grade book before I even opened the assignment. Javi wrote the whole thing on his phone during bus rides to soccer practice, and it had all the messy, specific details AI would never make up—he mentioned his grandma’s terrible burnt chocolate chip cookies and the way the bus smelled like old sneakers. The flag was wrong, but he still had to spend two afternoons meeting with admin to get it overturned.
The first thing I tell parents who reach out about this is that most AI detection tools aren’t actually very good at what they do. They’re trained to spot patterns they associate with AI writing: consistent tone, clear structure, lack of random tangents, and polished grammar. That means a lot of original student writing gets flagged for no reason, especially when kids use common tools to help them write.
Mia has dyslexia, for example. She uses text-to-speech to read her writing back to herself so she can catch awkward sentences, and she uses Grammarly to fix the spelling and punctuation errors she misses when she’s typing fast. That makes her writing cleaner and more consistent than it would be if she wrote it by hand, and that’s exactly what the detector reads as AI. Javi was a strong writer who planned his essay out carefully before he started, so his structure was clear and logical—another red flag for the detector. Even kids who use AI just to brainstorm a list of topics or organize an outline, then write the whole thing themselves, can end up flagged because the core structure matches a common AI pattern.
After dealing with this for a couple of years, both with my own kid and my students, I’ve picked up a few small, low-fuss things that almost always fix the problem. None of it requires fighting the school board or rewriting your kid’s paper for them, just a few extra steps added to the usual homework routine.
The first thing I tell everyone is to leave a paper trail of your process. Google Docs saves every version of your work automatically, but most kids don’t know that, and they’ll often delete old drafts or start a new document when they rewrite to keep it clean. I have Mia keep everything—messy first drafts, bullet point notes, even the random tangents she cuts out mid-writing. If she jots notes in her math notebook during study hall, she snaps a photo of it and saves it to the same folder as her assignment. When a flag pops up, the first thing we send is a link to the version history that shows her writing the paper line by line over three nights, plus the photos of her handwritten notes. No teacher I know will ignore that kind of proof.
Second, if you use any tool to help with writing, just disclose it upfront. A lot of kids think they have to hide that they used Grammarly, or that they asked AI to help them brainstorm topics, because they think it counts as cheating. But most teachers I know don’t care if you use a tool to help with the busy work, as long as the ideas and the final writing are yours. I have my kids add one line at the end of their assignment that says what tools they used: “I used Grammarly to correct spelling errors and text-to-speech to proofread, all analysis is my own.” It takes 10 seconds to write, and it makes a teacher way more likely to dismiss a flag as a false positive before they even start investigating.
Third, if you do get flagged, ask for a 10-minute oral check-in instead of just arguing over email. If a kid wrote the paper themselves, they can answer basic questions about their argument, their process, and how they came up with their ideas. I’ve never had a student who wrote their own paper stumble through that check-in, and it’s way faster than digging through all the paperwork to prove your case.
Last week, Mia got flagged again on her biology lab report. We sent the version history and her little disclosure note about using text-to-speech, and her teacher cleared it before the end of the school day. It’s still an extra step none of us should have to take. Kids should be able to turn in a paper and have it graded on its own merits, not whether an algorithm thinks it sounds “real enough.” But until schools stop relying on these janky tools, this is the system we’re working with. These small steps don’t fix the root problem, but they cut down on the 8:30 p.m. kitchen table fights, more often than not. It’s not perfect, but it works for us right now.
Please indicate: Thinking In Educating » Why Your Teen’s Assignment Keeps Getting Flagged for AI (What I Do As A Teacher Mom)