What Should Kids Learn In the Age of AI? Practical Tips for Parents
Last Tuesday, I sat across from my 12-year-old niece Lila at my kitchen table, half a bag of cheddar Goldfish crumbled between us, the dishwasher humming in the background. She prodded her laptop screen, where a 7th grade science lab report prompt glowed: Explain the chemical reaction of your baking soda volcano experiment. “Can I just get ChatGPT to write this?” she said, not looking up. “All my friends do.”
I almost said no, out of habit. The old rule was you do your own work, no cheating. But my brother, Lila’s dad, is a backend developer, and he’d told me just two nights before that half his job these days is editing AI-generated code, not writing it from scratch. So I leaned forward and asked, “What would you do after it writes it?” She shrugged, picked a Goldfish crumb off the table. “Change a few words so the teacher doesn’t notice.”
That question has been bouncing around my head ever since. All over social media and parent email chains, people are arguing about what kids need to learn now. Some say everyone has to learn coding. Others say the future is all STEM, so drop art and music. But half the entry-level coding jobs I see posted already ask for experience working with AI, not just writing code from scratch. So what do we actually tell our kids to focus on?
Two weeks before Lila’s homework question, I’d sat in my 10-year-old son’s elementary school gym for parent-teacher conferences. His art teacher, Ms. Henderson, pulled me aside to show me two projects. One was a perfectly polished forest scene, generated by AI, printed and turned in by a 5th grader. The other was a messy, lopsided drawing of a cat with three ears, done by a 2nd grader. Ms. Henderson pointed to the messy one. “This kid could tell me why the cat has three ears – it’s her grandma’s cat, who lost an ear to cancer, and she drew it how it really looks. The AI kid couldn’t tell me a single thing about why he picked a forest, or what he liked about it. Who do you think is learning more?”
I told Lila she could use AI to get a first draft, but she had to do three things before she turned it in. First, cross out every line that didn’t match what actually happened when she did her experiment. AI didn’t know that she used too much vinegar the first try and got baking soda all over her mom’s kitchen tile. It didn’t know that she found out half a cup of baking soda worked better than the two cups AI recommended. I made her add those parts, and write one paragraph explaining what that mistake taught her that the generic AI explanation didn’t cover.
That’s the first thing I think we need to tell kids to practice: putting what they think and what they’ve seen into their own words. AI can write anything, but it can’t write your experience. It can’t tell the story of the mistake you made that changed how you think about a problem, or the small detail that matters to you even if it doesn’t fit a generic answer. Practicing that takes five minutes a day, just asking your kid to tell you about what they did at recess that day, or why they liked that movie, instead of letting them shrug and say “it was fine.”
The second thing is learning how to ask questions that get you the answer you actually want. A couple of weeks ago, my son Sam was working on a project about bees for his science class. He typed “tell me about bees” into AI and got a generic three paragraphs about pollination and life cycles. It was boring, and he almost turned it in. I sat with him for five minutes and had him rephrase it. We ended up with “why do bees love the clover in my backyard more than the rose bushes in my neighbor’s yard?” That question got him specific, useful information, and it made him curious enough to go outside and sit on the grass for an hour counting bees, just to check if it was true.
AI can only work with what you give it. If you ask vague questions, you get vague answers. Teaching kids to add specific details to their questions, to ask for what they actually care about, is a habit that sticks. I started asking Sam to rephrase generic questions when he asks me for help, and he does it on his own now more often than not.
Third, practice spotting what’s wrong. AI makes up facts all the time – it’s just guessing what word comes next. I tell my kids that any time they get information from AI, they have to confirm it with at least two other sources. If AI says baking soda volcanoes need two cups of baking soda, check your experiment, check a textbook, check a trusted science website. If it says a certain historical event happened on a specific date, double check. It doesn’t take long, it’s just a habit. That habit is way more useful than memorizing a bunch of facts that AI can pull up in two seconds.
Last month, I was talking to my neighbor Joe, who owns a small landscaping business here. He hires three or four high school kids every summer. He told me he fired one kid last year who just used AI to draw up garden plans for customers. The plans looked great on paper, but the kid couldn’t adjust when a customer said they didn’t want to spend two hours a week weeding, or they needed a plan that was safe for their dog. The kid he kept, the one who comes back every summer, can listen to a customer ramble for 10 minutes, pull out what they actually want, and adjust the AI plan to fit. That’s the skill that got him a part-time raise after his first summer. AI can’t read the offhand comment a customer makes about hating yard work, not yet anyway.
Lila turned in her lab report last week. Her teacher marked off a few points for using AI to draft it, since the class rule is first drafts have to be student-written. But she gave Lila extra credit for the section about her messy first eruption and the adjustment she made to the recipe. We still don’t know how all this AI stuff is going to shake out. Schools are still rewriting rules, I’m still changing how I talk to my kids about homework when I see new things pop up. No one can say for sure what jobs will exist when my kids are my age. But so far, these small, boring habits – explaining your own experience, asking specific questions, checking for mistakes, listening to what other people actually mean – seem way more useful than any of the fancy future-proof skills people are selling parents these days. It’s not perfect, and we’re all figuring it out as we go. But it’s a start.
Please indicate: Thinking In Educating » What Should Kids Learn In the Age of AI