Is AI Making Students Worse At Sitting With Confusion? What One Middle School Teacher Noticed
The projector hummed over the crinkle of a Goldfish bag and the scratch of pencils on graph paper last Tuesday, halfway through my 8th grade Algebra 1 block. Mia, who sits in the front row with a glittery water bottle and usually turns her work in five minutes early, raised her hand before most of the class had even finished reading the word problem. I walked over, and she tilted her notebook toward me, half-frustrated, half-exasperated. “Can you just tell me if I’m on the right track? I don’t get this, and I’ll just pull it up on ChatGPT otherwise.”
I’ve been teaching middle school math for 12 years, and this line feels far more common now than it did even three years ago. Five years ago, if a kid got stuck, they’d slump back, chew on their eraser, cross out three lines of work, turn to their neighbor and complain, then try again. Now, the second that twist of confusion hits their chest, they reach for their phone. It’s not just AI, of course—we’ve been moving toward instant answers for decades with search engines and answer apps—but AI has made getting a complete, perfectly polished answer faster and easier than ever. That convenience comes with a tradeoff I’ve watched play out in my classroom and in parent conferences over the past year: kids are losing practice sitting with the discomfort of not knowing.
A couple of weeks back, I assigned an open-ended project: build a budget for a hypothetical class field trip, with no single right answer, just a requirement to justify every choice. Half the class turned in perfectly formatted, error-free budgets. When I called on kids to explain why they’d budgeted $12 per person for admission, most just shrugged and said “AI came up with that.” They hadn’t spent any time confused about whether we should cut snack costs to afford a better museum ticket, or whether the bus price I gave matched the number of kids we had. They skipped all that messy thinking entirely.
After class, I met with Mia’s mom for a fall conference, and she brought up the same pattern at home. Mia used to spend Sunday afternoons doing crosswords and 500-piece jigsaw puzzles with her grandma. Now, if she can’t get a clue in 10 seconds, she pulls up the answer app. If a section of the jigsaw doesn’t fit right after five minutes of trying, she quits and scrolls TikTok. It’s not just school. It’s any time something doesn’t click immediately.
I don’t blame kids, and I don’t think AI is some evil force that’s ruining learning. I use AI myself to generate extra practice problems and sort permission slip forms; it’s a great tool that cuts down on my busy work. But what I’ve started to notice is that we’re treating confusion like a problem to be solved as fast as possible, instead of what it actually is: the core part of learning. When you sit with that confusion, your brain digs through old stuff you’ve learned, connects it to the new problem, and stretches. If you yank out the answer immediately, you cut that process off before it even starts.
Over the past semester, I’ve tested small, low-effort changes to help kids rebuild that habit, and most of them are things parents can easily do at home too. First, I made a non-negotiable rule: 15 minutes of no AI, no phones, for any problem-solving work. Before anyone reaches for a tool, they have to write down three things they do know about the problem, and one thing they’re confused about. It takes two minutes, and it forces them to engage instead of bailing immediately. At home, this translates to something as simple as keeping a small basket by the homework table, where all phones go for the first 15 minutes of homework time. For hobbies or puzzles, it’s just 10 minutes of trying before you check an answer.
The second small shift I’ve made is in what I say when kids say they’re stuck. Instead of saying “just use AI” or “figure it out yourself,” I say, “What’s one small step you can try right now?” It’s a tiny change, but it gets them moving instead of shutting down. Last week, after I reminded Mia of the 15-minute rule, she wrote down her three things: I know we’re looking for the number of adult and student tickets, I know the total number of tickets is 120, I know adult tickets cost $15. The thing she was confused about was how to set up the second equation for total cost. She stared at it for three minutes, scribbled something out, and poked me again: “Wait, is this it?” It was. She’d figured it out herself, and she got that little rush of a win you can never get when AI hands you the answer.
It’s not that I ban AI entirely. After that 15 minutes of working through confusion, kids can use it to check their work, or to walk through a step they still don’t get. I just want them to do the messy, uncomfortable part first. I’ve noticed that since I started the 15-minute rule, kids talk to each other more. Last week, two boys in the back row spent 10 minutes arguing about a probability problem, going back and forth, before they figured out they’d both misread the question. That kind of back-and-forth, that messy confusion, that’s where the learning sticks.
We’re still figuring this out. Some kids still sneak their phones into their hoodies, some parents tell me they don’t have time to wait 15 minutes of staring at a homework problem when there’s soccer practice and bedtime to get to. I get that. This isn’t about being perfect. It’s just about remembering that that uncomfortable feeling of not knowing isn’t a sign you need to get an answer right away. It’s a sign your brain is working. Right now, we’re just trying to build a little more patience for that feeling, one problem at a time.
Please indicate: Thinking In Educating » Is AI Making Students Worse At Sitting With Confusion