Why the push for AI-era critical thinking skips what students actually need
Last Wednesday, I was sitting at my scarred 10th grade English desk, sipping lukewarm coffee from a mug covered in my 7-year-old’s dinosaur sticker, grading argument essays on the prompt “Should public schools ban student use of AI?” Most essays hit the same beats we’d drilled in our recent unit: AI produces misinformation, it amplifies bias, students need to question every output it generates. Then I got to Mia’s.
Mia’s essay was perfectly structured. She cited three studies on deepfake accuracy, broke down how AI chatbots mirror the political leanings of their training data, and concluded that schools need stricter critical thinking lessons to counter AI’s harms. Tacked on at the end, in tiny parentheses below the works cited, was one line: “I don’t actually care that much either way. No one asked me what I would use AI for.”
I stared at that line for a few minutes, then packed up to wait for after-school help, like I do every Wednesday. Three students lingered after the bell that day, and when I mentioned Mia’s line, all three started talking over each other. Javi, who lives with his single mom working night shifts at the local hospital, said he uses ChatGPT to walk him through algebra problems he doesn’t understand when his mom isn’t home to help. Lila, who gets anxious about raising her hand multiple times in one period, said she uses it to brainstorm story ideas when she gets stuck, so she doesn’t have to bother me five times in 20 minutes. None of them mentioned cheating. None of them talked about misinformation. They talked about what AI does for them, what gaps it fills that school and busy families can’t always cover.
That conversation got me thinking about the professional development we did back in August. All day, presenters talked about the urgent need to teach critical thinking for the AI age. We did activities on how to spot edited images, how to cross-check AI-generated facts, how to push back on flawed AI arguments. I left feeling motivated, adjusted my unit plans, added all the new drills. It made sense, on paper. But what I’ve noticed over the last two months is that all that training starts from one unspoken assumption: that kids already know what questions they want answered, what problems they want to solve. All they need to do is critique the answers AI gives them.
But that’s not what I see in the classroom every day. Most of my students can tell you what argument is expected of them on a test or an essay. Most can spot a bad source if you tell them which source to check. But far fewer can tell you what they care about, what they actually want to know, beyond what’s on the rubric.
Two weeks ago, I tried a small shift for our new climate change unit. Instead of opening with a critical thinking exercise on fact-checking AI-generated climate claims, I put a blank slide up and asked everyone to write down one question they have about climate change that they’ve never gotten to ask in school. No wrong questions, no off-limits topics. The first question I got was from a freshman: “Can I still get a golden retriever puppy if dog food is bad for the environment?” Another: “Will my grandma’s house in Florida be gone before I graduate high school?” Another: “What job can I get that actually helps fix climate change instead of just making money?” None of these were the polished, argument-ready questions I usually put on the board. None of them lined up with the unit test I’d planned. But every kid wrote something. No one turned in a blank paper, which almost never happens on the first day of a new unit.
The shift here isn’t that we need to throw out critical thinking lessons. It’s that we need to make space for the step that comes first, that almost everyone is skipping right now. It doesn’t take a big curriculum overhaul or expensive new training to do this. For parents, when your kid sits down to do homework and mentions they used AI to help, instead of leading with “Did you fact-check that?” try asking “What do you still want to find out that you haven’t gotten yet?” It’s a small shift in the order of questions, but it changes the whole conversation. For teachers, leave 10 minutes at the start of every new unit for kids to write down their own unfiltered questions, and pick one or two to explore as a class that week, even if it means pushing a lesson plan back a day. Last week, we spent two full days talking about pet carbon footprints, and half the kids stayed after class to share what they’d found on their own. That’s more engagement than I’ve gotten from a full month of pre-planned critical thinking drills.
Last Friday, Mia stayed after class to talk about her essay. She twisted that same silver hair tie around her wrist the whole time, the one with the little cat charm that’s fallen off one end. She told me she has generalized anxiety, and she freezes up when she doesn’t understand a practice test question, and she’s too embarrassed to ask the same question three different times in class. AI lets her go over it as many times as she wants, no judgment. She wrote the first essay the way she knew I wanted it, she said, but she wanted to write about why schools should let kids use AI as an accessibility tool for kids like her instead of just banning it or teaching them to critique it. I told her that was fine, she could turn in whatever she wanted.
She’s still working on it. I don’t know if it’ll have all the correct citations or hit all the critical thinking points we’ve been teaching. But when she walked out of my classroom, she was actually smiling about an essay, which is something I haven’t seen from her all semester. We’re still figuring this out. We still need to teach kids how to spot bad information and question flawed arguments. But right now, the most important thing we can do is stop jumping straight to critiquing output, and start asking kids what their own questions are first. It’s messy, it throws off lesson plans, it doesn’t fit neatly into a standardized test question. But so far, it’s the most useful change I’ve made all year.
Please indicate: Thinking In Educating » Why the push for AI-era critical thinking skips what students actually need