Is AI helping kids learn or just helping them avoid hard work? What I’ve seen as a 4th grade teacher
Last Tuesday, during math workshop, I knelt next to Javi’s desk to check in on his multi-step word problems. His fingers flew to the bottom of his Chromebook screen, minimizing a tab before I could see what it was. Javi’s got a hole in the elbow of his grey Minecraft hoodie, and he always twists the end of his pencil when he’s nervous. “It’s not what you think,” he mumbled, but I already knew. I let my class use AI for extra help when they’re stuck, but he’d gone and typed the whole problem set into the chat box and copied the answers. When I asked him why, he shrugged and said, “It’s just confusing. Why do I have to do all the steps when it already gives me the right answer?”
That same question was bouncing around my head when I got home that night and checked on my 13-year-old daughter Lila, who was cramming out a history term paper about the Great Depression. She had AI pulled up on her laptop, and I watched her tweak a sentence here and there in a full outline the tool had written for her. When I asked her what she found most interesting about her topic, she blinked and said, “I dunno, I haven’t really thought about it yet. AI did the outline already.”
That’s when I started really leaning into the question I hear from so many parents and teachers lately: are we using AI to help kids learn, or are we just letting them skip all the hard parts that actually make learning stick? I used to write AI off as just a new version of the same old cheat sheet kids have been hiding in textbooks forever, but after a few months of watching it in my classroom and my own home, I’ve learned it’s not that simple. It can be a really helpful tool, but it’s way too easy to let it do the work that actually builds thinking skills.
The tight, frustrated stuck feeling you get when you’re trying to untangle a problem or sort out what you actually think about a topic isn’t a sign you’re bad at learning. It’s the point of learning. When I spend 20 minutes stuck on a tricky bread recipe, I don’t just remember the recipe better next time— I learn what adjustments to make for my drafty old oven. Same thing with kids. When Javi spends 10 minutes figuring out which step comes first in a word problem, he’s not just solving that one problem. He’s learning how to work through confusion, which is the skill he’ll need for every hard thing he does after 4th grade—getting a driver’s license, managing a budget, fixing a leaky faucet. When AI does that part for him, he doesn’t get that practice.
I don’t ban AI in my classroom, and I don’t ban it at home. Banning it doesn’t make sense— kids are going to use it for the rest of their lives, for school and work, so they need to learn how to use it right now instead of sneaking it behind our backs. Instead, I’ve landed on two simple, low-fuss rules that work both for my 28 4th graders and my own teen. The first rule is: write down your thinking before you open AI. In my class, that means before any kid can ask AI for help, they have to write down at least one thing they already tried, and one specific thing they’re confused about. They can’t just ask AI to solve the problem. They have to write “I know we need to find the total cost of the field trip, but I don’t know if I need to add the discount before or after the bus fee” before they pull up the tool. Then, they have to ask AI to explain the step they’re stuck on, not give them the whole answer. At home with Lila, that means she has to hand-write three main points she cares about making in her paper before she uses AI to help her organize it or find extra sources she missed.
Last month, Lila was working on her science fair project about how different liquids affect bean plant growth. Her entire control group, the ones watered with plain tap water, all died after a week, and she couldn’t figure out why. She sat twisting her sour candy wrapper at the kitchen table for 20 minutes, frustrated, before she asked if she could use AI for help. She wrote down what she’d done: same amount of water every day, same spot by the window, only difference was the liquid. She wrote “I think maybe the tap water has too much chlorine? Or did I overwater them?” Then she asked AI what common mistakes new gardeners make with bean seeds. It turned out she’d planted the seeds twice as deep as the package said. She fixed it for her second trial, adjusted her conclusion, and ended up taking third place at the school fair. AI helped her get unstuck, but she still did all the thinking about her own experiment, and she still learned how to troubleshoot when something goes wrong.
It doesn’t work 100% of the time, of course. Last week, I caught Javi cutting corners again during independent work. He’d written down a fake stuck point just so he could use AI to get all the answers fast so he could play the math game we let kids do when they finish early. I didn’t give him a detention or send a note home. I just pulled a chair next to his desk and said “Let’s do this one together, step by step.” Halfway through the third problem, he stopped twisting his pencil and said “Oh, I get it now. That wasn’t as hard as I thought it’d be.” He just forgot that the stuck feeling doesn’t last forever.
I still don’t have all the answers. Some weeks I think we should just put the Chromebooks away and do all work on paper, and some weeks I see AI help a kid who’s been stuck for days stop feeling stupid and keep going. Right now, the rule of “do the hard thinking first” is working for us. It doesn’t stop all the cutting corners, but it gives kids a chance to actually learn, while still getting the help AI can offer. I don’t know what school and work will look like for these kids in 10 years, but I know they’ll still need to know how to work through hard, confusing things on their own. That’s one thing AI can’t do for them.
Please indicate: Thinking In Educating » Is AI helping kids learn or just helping them avoid hard work