Why Your Kid’s Assignments Keep Getting Flagged for AI (And What We Did About It)
It’s 8:30 on a Tuesday, the kitchen table still smells like the garlic bread we had for dinner, and my 10th grader Javi just slammed his laptop shut hard enough to make his half-eaten bowl of mac and cheese jostle. “It happened again,” he says, jaw tight. He yanks the laptop open and spins it toward me. Across the top of his 4-page New Deal essay, a bright red banner blares: 92% AI probability detected.
This is the third time in five weeks. The first was a bio lab report back in September, the second a poetry response for English, now this. Javi’s been playing varsity soccer this fall, so he’s been squeezing homework in between 6 am practices and 7 pm game days. This essay took him three nights of work after cleats came off, and he’s furious that all that work’s suddenly in question. The first time we got a flag, I’ll admit I hesitated before I asked him straight if he’d used a chatbot to write it. He got so upset he pulled up every tab he’d had open, every handwritten note he’d jotted in study hall, to prove he’d done it himself. I believed him, but it got me wondering why this kept happening to a kid who’d rather forget his lunch than cheat on a test.
I emailed his history teacher Ms. Henderson that night, and she called me back Thursday while I was folding Javi’s stinky soccer laundry. She didn’t even sound surprised. The school forced all teachers to turn on these AI detectors after a group of seniors cheated on their AP Bio final last spring, she said, and she gets 10 to 15 false flags a week. Eight out of 10 of those flags come from kids who didn’t touch AI at all.
She explained it simply: the detectors are trained to spot what they think “human” writing looks like—messy, a little rambling, full of tangents and awkward phrasing. When a kid writes a clear, structured essay that sticks to the prompt, follows the thesis statement they teach, and cuts out all the extra fluff that doesn’t add anything? The algorithm thinks that’s too polished to be a teenager. It doesn’t help that Javi has ADHD. He’s spent two years working with a tutor to learn to cut his tangents and stay on topic, because he used to turn in great work that wandered off to talk about WWII documentaries or new soccer cleats and got marked down for it. Now that he’s doing exactly what every teacher has told him to do, he’s getting penalized for it.
We talked through what we could do, and she gave us the first tip that’s worked ever since: share your full edit history. Almost every kid uses Google Docs for school now, and it saves every single change you make from the first word you type. All Javi has to do is turn on link sharing for the draft, drop the link in the submission comments, and any teacher can see exactly how the essay came together. For Javi’s New Deal essay, Ms. Henderson could see he started drafting six days before it was due, added a paragraph after he watched a documentary in class, deleted an entire section on public works when he decided his argument was stronger focusing on Social Security, and changed his thesis twice. No one can fake that level of incremental change if they just generated the whole essay in two minutes on a chatbot. The flag was cleared before the end of the day.
We added a second small step after that: a 3 to 4 sentence writer’s note at the end of every assignment, ungraded, just a quick breakdown of Javi’s process. For the New Deal essay, he wrote: “I got stuck for an hour trying to figure out why Southern Democrats blocked universal healthcare back then. I ended up going with what the textbook said because I ran out of time, but I want to talk about it more after class next week.” It’s messy, it’s honest, and it’s obviously a kid talking, not an algorithm. Ms. Henderson told me she actually looks forward to these notes now—they give her a better idea of what students actually understand than the polished essay itself.
When we hit a snag with a new bio teacher who wouldn’t look at edit history and insisted any flag meant an automatic zero, Javi just asked for a five minute check-in after class. He walked her through his lab report step by step, explained how he collected his data, what he changed from his first draft, and why he’d messed up his first calculation. She cleared the flag within ten minutes. It took less time than arguing back and forth over email, and it’s pretty hard to fake that level of detail if you didn’t do the work yourself.
I used to assume that any flagged assignment meant the kid was cutting corners. Now I know how broken these tools are. They punish kids for learning good writing habits. They penalize kids with learning differences who’ve worked extra hard to clean up their work. They add unnecessary stress for everyone—kids, parents, teachers—who just want work graded fairly.
Last week, Javi turned in a book report on To Kill a Mockingbird and got flagged again. He sighed, added his writer’s note, dropped the edit history link into the comments, and went back to playing Mario Kart with his friend. The flag was cleared the next day. It’s still annoying. It’s still an extra step he shouldn’t have to take. The school is still talking about changing their policy, but nothing’s shifted yet. For now, this is what works. It’s not a perfect fix, but it keeps his grades from getting hurt for something he never did.
Please indicate: Thinking In Educating » Why Your Kid’s Assignments Keep Getting Flagged for AI (And What We Did About It)