Why the Push for AI-Era Critical Thinking Can Overlook What Students Actually Need
Last month, I stayed after school to grade 7th grade argument essays, and Mia hung around by my desk, scuffing her soccer cleats on the tile, waiting to talk about her B. The prompt was “Should schools ban student use of AI chatbots?” and my feedback at the top read “Needs more critical analysis of AI’s harms – strengthen your pushback against unregulated use.”
She twisted the hem of her jersey, chipped blue nail polish catching the hum of the dying fluorescent light. “I don’t get what pushback I’m supposed to give,” she said. “My aunt had a stroke last year, right? Her hands shake too bad to write. She uses ChatGPT to draft birthday cards for my grandma, then she copies them out slow, and it’s the only way she can do it anymore. What am I supposed to say? That’s bad?”
This year, every professional development session I’ve sat in has centered the same priority: we have to teach AI-era critical thinking. Kids need to spot deepfakes, fact-check hallucinations, question algorithm bias, push back against AI-generated arguments. All of that is true. But after talking to Mia, I started counting how many times I’d asked students to critique something this semester, versus how many times I’d asked them what they actually thought, rooted in the things that have happened to them. The gap was huge.
Two weeks later, I graded a 10th grade essay on teen social media use from Javi. It was perfectly polished, hit every critical thinking benchmark I’d laid out. He cited three peer-reviewed studies, broke down how algorithmic amplification worsens body image issues, pointed out gaps in recent research that other students missed. When I pulled him aside after class to ask how he’d gotten interested in the topic, he shrugged. “My mom won’t let me have TikTok. I just pulled all the stats from an AI outline. It’s what you wanted, right?” He didn’t have a single personal line in the whole three pages, because I’d never asked for one. I’d only asked for critique.
My sister, a parent of a 9-year-old, texted me about this same pattern last week. She’d been reading all the parenting columns about teaching kids to be critical of AI, so every night when her son does his homework, she quizzes him on what the AI got wrong, what biases it has, why he can’t trust it. “He used to ramble for 20 minutes about whatever dinosaur he was studying that day,” she said. “Now he just says ‘this AI got the T-Rex timeline wrong, can I go play?’ He won’t talk about the things he cares about anymore.”
We’re so scared of kids accepting bad AI output at face value that we’ve jumped straight to teaching critique before we help them form their own beliefs in the first place. We assume critical thinking is a skill you can pour into an empty head, but it doesn’t work that way. You can’t push back on an idea if you don’t first have a sense of what you believe, rooted in the things you’ve seen and lived. That’s the part we’re overlooking right now.
It’s a small shift I’ve started making in my class, and my sister has started trying at home, and it doesn’t take any extra time. Now, when we start any new topic – AI, social media, even a novel with a controversial theme – I spend the first five minutes just collecting personal stories, no judgment or critique allowed. Instead of opening with “What are the risks of deepfakes?” I ask “Has anyone ever had someone share something that wasn’t true about you, or a friend, online? Tell us that story first.” For parents at dinner, that can be as simple as swapping “Did you learn how to spot bad AI at school this week?” for “Has AI helped you with something this week, or bugged you? Tell me about that time.”
Another small change I’ve made: when I grade now, I look for a personal detail first, before I check for critical thinking gaps. If there isn’t one, I don’t mark down for bad analysis – I write back “Add one thing that happened to you that makes you think this, and we’ll talk about the rest next.” My sister does something similar: when her son brings home homework he used AI to help with, she first asks what part he thought was the coolest, before they talk about what the AI got wrong.
I changed Mia’s grade to an A- after our talk. I told her the line she’d tossed off at the end – “AI isn’t good or bad, it just helps people do things they can’t do anymore” – was the most interesting thing in the whole essay, and I wanted her to build the rest of the argument around that. She came back the next day with two new paragraphs: one about the birthday cards, and another about the time a girl in our grade used AI to make a fake meme of another student that got shared all over the school. She came to that critique on her own, because she started from what she knew, not what she thought I wanted her to say.
I still teach all the critical thinking lessons. I still walk kids through how to spot AI hallucinations, how to question bias, how to fact-check what they read. I just start 10 minutes later than I used to. Some days, it feels messy. We don’t always get through all the slides I planned, and there are days I worry I’m not checking enough boxes for the district’s new AI literacy requirements. But the essays I get now don’t read like generic takes with extra critique tacked on. They read like kids thinking, for themselves. That’s the thing we can’t afford to skip, no matter how urgent the push to prepare kids for the new technology world feels. It’s still working so far, but I’m still figuring it out as I go.
Please indicate: Thinking In Educating » Why the Push for AI-Era Critical Thinking Can Overlook What Students Actually Need