What Happens When AI Removes the Bottom Rungs of the Learning Ladder
Last Tuesday, I sat at my kitchen table after school, a half-eaten bowl of buffalo mac and cheese gone cold next to my chipped ceramic coffee mug, grading 8th grade algebra homework. Javi Mendez, the kid who wears a faded Arsenal soccer hoodie even when it’s 80 degrees out, turned in a perfect set of 15 slope-intercept problems. Every line of work was neat, every answer correct. That would have been great news, except I’d pulled Javi aside for extra help just last week, when he still couldn’t consistently convert problems from standard form to slope-intercept without mixing up his signs.
I pulled him aside before homeroom the next day, leaning against the cinder block wall by my classroom door. “How’d you get every single one of these right?” I asked. He shrugged, pulled his phone out of his pocket, and pulled up a ChatGPT window with all the problems pasted in. “I just plug ’em in, it shows the steps. I copy it down. Why would I do all that multiplying and dividing myself when it does it in 10 seconds?”
I didn’t yell, didn’t take his phone, just nodded and sent him to homeroom. But I kept thinking about last fall, when Javi stayed after class for 20 minutes, crumpling up three papers because he kept messing up that same sign conversion. We went through it slow, one problem at a time, and he finally got three right in a row. He pumped his fist under his desk and grinned, like he’d just won a soccer game, not done middle school math. That day stuck with me, and when Javi showed me his phone, I got what people mean when they talk about AI pulling out the bottom rungs of the ladder.
I saw it play out at home two weeks later, too. My 12-year-old daughter Lila had a school spelling bee coming up, and she complained for 20 minutes about having to write each word 10 times to memorize them. She said she’d found an AI tool that makes interactive multiple-choice quizzes for free, no writing required. I shrugged and let her try it—why not, if it works? She got out in the second round on “separate,” spelling it “seperate,” the exact mistake she would have avoided if she’d written it out enough times to feel the order of the letters in her hand.
The boring, small, repetitive steps no one likes are the bottom rungs. They’re spelling practice, basic fraction problems, writing out topic sentences by hand, all the stuff that feels like a waste of time when you can get AI to do it faster. We love celebrating shortcuts, we love skipping the work that feels like drudgery. But if you pull the bottom rungs out, there’s nowhere for your feet to go when you want to climb higher. You can read AI’s explanation of algebra, you can even memorize the steps it shows you, but if you never did 20 basic problems yourself, you don’t have the intuition to spot when you’ve messed up, or to adjust when a test problem is a little different than the one AI solved for you.
I’m not anti-AI. I use it myself to cut down on the hours I spend making rubrics and formatting handouts, it’s saved me countless evenings I would have spent at my desk instead of with my kids. But I’ve changed small, concrete things in my classroom and at home to keep those bottom rungs in place, and they’ve worked better than I expected.
In my classroom, I don’t ban devices or AI. I just changed the homework rules: for any new skill we’re learning, the first 12 out of 15 practice problems have to be done by hand, no devices out on desks. Students can only use AI on the last three challenge problems, and only after they’ve tried them themselves first. If they use AI to get a hint for a hard problem, they have to close the app and write out the next three steps on their own before they look back. That takes 30 extra seconds to enforce, it doesn’t require any fancy new programs or training.
At home, we keep a woven basket by the front door, and all phones go in it for the first 30 minutes of homework time. That’s when we knock out all the basic skill practice—spelling, math facts, early algebra problems—before we pull out any devices for help. If Lila uses AI to work out an essay outline, she has to write the full first draft of every body paragraph by hand before she lets AI help tweak wording or fix awkward sentences. It’s a small habit, no big fight, just a routine we stick to.
After our talk by the classroom door, Javi started staying after once a week to do the basic practice he’d skipped the first few weeks of the unit. Last week, he did 10 conversion problems on his own, no hints, no phone, got nine right. He leaned back in his chair, wiped his forehead on the sleeve of his soccer hoodie, and grinned that same fist-pump grin I saw last fall. It wasn’t a perfect test score or a award, just the quiet feeling of getting it right yourself.
I don’t think AI is going anywhere, and I don’t think we should try to pretend it doesn’t exist. There are good uses for it, for teachers and for kids. I still don’t have all the answers, some days I wonder if I’m fighting a losing battle to keep slow, boring practice relevant. But then I see Javi’s grin, or Lila aces her spelling test after writing all her words out by hand, and I know it’s worth the extra 10 minutes a day. You can’t jump from the ground to the third rung and expect to stay up. The bottom rungs are still there, even when it’s easier to pretend you don’t need them.
Please indicate: Thinking In Educating » What Happens When AI Removes the Bottom Rungs of the Learning Ladder