Why My Middle Schooler Hates Mandatory AI Assignments (And What We’re Doing About It)
The Chromebook hit the kitchen counter hard enough to make Lila’s half-eaten cheddar Goldfish skitter across the worksheet underneath. It was 3:47 on a Tuesday, leftover tomato soup congealing in the pot on the stove, her 7-year-old brother dumping Lego bricks on the floor by the fridge, and my 13-year-old was already declaring she’d take a zero on her history essay before she did a single minute of work. “They’re making me do it again,” she said, crossing her arms. “First step, paste the Oregon Trail letter into ChatGPT, get three core themes, then build your argument from that. If I don’t turn in a screenshot of the AI output, I get 10 points off automatically. I hate it. I hate all of it.”
My first reaction, like most parents I know, was the line every adult has repeated for the last two years: “It’s just a tool, kiddo. Like a calculator. You have to learn to work with it, it’s not going anywhere when you get a job.” She just stared at me, then flipped her Chromebook around to show me the rubric. Twenty percent of the assignment grade was explicitly “effective use of AI for initial brainstorming.” She’d tried a workaround the month before, done all her own brainstorming first then run it through AI just for the screenshot, and gotten a note in the margin: “Need to use AI to generate starting ideas, not just check your own work.”
It’s not just history. Math now requires an AI check screenshot for every homework assignment, no exceptions. English vocabulary requires AI-generated sample sentences before students write their own. Even art class last week had them use AI to generate three reference sketches before anyone could pick up a pencil. She doesn’t hate AI entirely, I learned that once I stopped lecturing long enough to listen. She uses it on her own time to debug code for her after-school robotics team, and to look up 19th century slang when she’s reading old letters for fun. What she hates is being told she has to turn over the first, messiest, most interesting part of thinking to a machine before she gets to try it herself.
I sat down with her that night after her brother went to bed, no Goldfish crumbs, no Lego, just the two of us and the Chromebook, and told her I’d work through the assignment both ways with her, just to see what she meant. The primary source was a four-page letter 16-year-old Elizabeth Ann Gray wrote to her cousin back in Illinois in 1853, halfway through her family’s trip on the Oregon Trail. Lila did her pass first, scribbling notes on a scrap of notebook paper, and inside two minutes she’d circled three throwaway lines I never would have noticed: Elizabeth wrote about missing the old apple tree in her backyard three separate times, once between a line about a broken wagon wheel and another about a child in the next wagon dying of fever. When we pasted the full letter into the AI tool, it spit out exactly three polished, predictable themes: the physical hardship of overland travel, cultural conflict with Indigenous groups, and optimistic hope for economic opportunity in Oregon. No apple tree. Not a single mention.
“See?” Lila said, prodding the screen with her pencil. “It doesn’t care about the thing she actually thought about when she was sitting by the campfire at night. It just says what every other essay about the Oregon Trail already says. If I start with that, I never would have looked for the apple tree in the first place.”
I’d gone into the conversation ready to argue about job skills, and I left realizing adults are talking straight past kids on this one. Most school AI policies are written by people who remember learning to write papers and solve math problems the old way, and see AI as an add-on skill to teach on top of that. But for kids like Lila, it’s not an add-on. It’s a gate they have to pass through before they’re allowed to do the actual thinking part of school.
We didn’t storm the administration office or write a big angry email. We worked out small, doable adjustments that don’t break school rules and don’t force her to give up the part of learning that matters. First, I stopped leading with “it’s just a tool.” That line makes kids shut down, because it dismisses the very real loss of the unscripted, curious part of school work most kids actually enjoy. Instead of asking a broad “why do you hate AI,” I asked one specific question: “Which part of being forced to use it bugs you most?” For Lila, it was never the tool itself. It was being graded on whether she used it, instead of whether her ideas were good.
Her standard workaround now is simple: for every assignment that requires AI output, she does all her own brainstorming, drafting, and problem-solving first, offline, in a beat-up spiral notebook she keeps next to her Chromebook. Only after she’s already worked through her own ideas does she open the AI tool, plug in the required prompt, and save the screenshot for submission. When the assignment allows, she adds one short note: one thing the AI got right that she missed, and one small, specific thing she noticed that the AI skipped.
She talked to each of her core teachers one by one after class, 30 seconds at a time, no audience, no big demand for an opt-out. Her history teacher wrote back that night saying the district requires every core class to integrate AI into every unit this year, and she’s still figuring out how to do that without killing kids’ curiosity. She loved the apple tree observation, told Lila her workaround was totally fine, and said she’s going to offer it as an option for the whole class on the next essay. Her math teacher was stricter about the required screenshots, but when Lila explained she learns more from working through wrong answers on her own first before checking, he shrugged and said as long as the screenshot is there at submission, he doesn’t care when she runs the problem through the tool. She still does all 20 problems on paper first, marks the ones she’s unsure of, only checks those, and writes a one-line note next to every wrong answer about what she messed up—forgot to carry a one, mixed up the slope formula—before she corrects it. Last test, she got a B+, up from a C- the first month of mandatory AI checks, when she was just changing her answers to match the tool before she thought through the work.
For teachers navigating district mandates they didn’t write, the low-lift shift that makes all the difference is simple: stop making AI use a graded requirement. If you have to include it in a unit, offer it as one option among three for brainstorming or editing: draw a mind map by hand, talk through ideas with a partner, or use AI to generate starting points. Let kids choose when to bring the tool in, if they bring it in at all. AI literacy doesn’t mean forcing every kid to use it for every step of every assignment. It means teaching them to judge when it helps and when it gets in the way.
Last semester, before the mandatory rules rolled out, Lila won the school’s 8th grade writing contest for an essay about a 1912 Lowell mill girl’s letter that mentioned a pet sparrow she kept in a cage by her loom. She spent an hour after school digging through the library’s digital archives looking up what kind of sparrows lived in Massachusetts mill towns that year, just for one throwaway line in her essay. She says if she’d had to run that letter through AI for brainstorming first, she never would have stopped on the sparrow line at all. It would have been smoothed out before she even started writing.
Last week, she turned in that required autumn poem for English, where the rule was AI writes the first draft, students revise it. She wrote her own first draft in her spiral notebook first, about raking leaves with her grandpa last fall and the half-frozen garter snake they found under a pile of maple leaves. She ran the generic “write a poem about autumn” prompt through AI, saved the required screenshot, typed her own poem below the AI output, and added one line at the top: “AI’s poem was nice, but it didn’t have the snake.” Her teacher gave her an A and a sparkly sticker that said “great original thinking.”
The district sent a back-to-school update last week that next year, every core class will have at least three mandatory AI assignments per quarter, and we don’t know yet what the fine print will say. She signed up for the after-school nature writing club last week, the one that meets in the library and only requires a notebook and a pencil, no Chromebooks allowed, and I paid her supply fee without a second thought. We don’t have a big grand position on AI, not really. She uses it when it helps her build her robot, I use it to translate smudged old French recipes my grandma wrote down. It’s useful, sometimes. But right now, at 13, the most important thing she’s learning at school isn’t how to get a polished answer fast. It’s how to notice the apple tree, and the sparrow, and the half-frozen snake in the leaves. We’re just going to keep making sure she gets to look for those first, before anything else tells her what she’s supposed to see.
Please indicate: Thinking In Educating » Why My Middle Schooler Hates Mandatory AI Assignments (And What We’re Doing About It)