What Should I Tell My Child to Learn in the Age of AI? Practical Everyday Tips
Last Saturday morning, I was folding clean soccer uniforms at the kitchen table, my half-empty coffee mug sweating next to the basket, when my 11-year-old Leo spun around in his chair and dropped the question on me. He was mid-way through a science project on native bee population decline, his laptop open to a chat window, a crumb of strawberry pop-tart stuck to the edge of his notebook. “Mom,” he said, “why do I even have to remember the parts of a flower? AI can tell me that in two seconds. What am I supposed to actually learn that AI can’t do for me?”
The golden retriever was chewing on a stray sock under the table, and I paused mid-fold, the stretchy polyester of his practice jersey pulled between my hands. I’d scrolled enough think pieces to know the hot takes: Some say all kids should learn to code. Others say creativity is the only thing that can’t be automated. But none of those felt like an answer I could give my kid over pop-tarts and laundry. I mumbled something about school requirements and sent him back to his project, but the question stuck with me.
Three days later, I picked Leo up from after-school woodworking club, and I saw the answer sitting on the workbench in front of him. It was a lopsided birdhouse he’d been working on for three weeks. The first time he cut the entrance hole, he measured wrong and it was too big for wrens, the birds he wanted to attract. He’d filled the gap with wood putty, sanded it down for 20 minutes after club two days in a row, and cut a new hole an inch to the left. His instructor leaned over to me and said half the kids quit when they mess up a cut these days. They pull up a perfect design on their phones, get frustrated when their real wood doesn’t match the clean lines on the screen, and just walk away.
That’s when it clicked for me. We don’t need to teach kids some fancy new skill to keep up. We just need to double down on three small, ordinary things that AI can’t do for them, and they can practice every day at home.
The first is how to spot when something is wrong. Leo pulled all his initial stats for the bee project from AI, and one line said local bee populations had dropped 70% in 10 years. That didn’t sit right with him, because he’d spent a weekend helping our county extension office count bees at the local park back in spring, and the volunteer coordinator said the number was closer to 20% in our area. I didn’t have to launch into a big lesson. He’d already gotten in the habit of checking one source against another, because I’ve made that a small routine for years: Whenever he pulls a fact from anywhere—TikTok, a textbook, AI—I ask, “Where else can you check that to make sure it’s right?” It’s not a grand academic exercise. It’s just a 30-second question we add to homework time. Last month, his group project for social studies went sideways when one kid turned in a full AI-generated essay that got all the dates wrong for our town’s founding. The other kids didn’t think to check, because they assumed AI was right. That’s the skill that matters more than anything right now: knowing that no tool is always correct, and you have to verify the output against what you know of the real world.
The second is how to fix your own mistakes. AI can generate a perfect birdhouse design, a perfect essay, a perfect recipe. But it can’t sand the wood putty when you cut the hole wrong. It can’t rewrite the essay when it doesn’t sound like you. It can’t run to the store for new eggs when you drop the cake on the floor. Most parents, me included, are used to stepping in to fix small mistakes to save time or frustration. But lately, I’ve been stopping myself. When Leo forgets his soccer cleats, I don’t drop everything to bring them to practice. When he messes up a cut on his birdhouse, I don’t re-cut it for him. When he writes a bad first draft of a book report that AI helped outline, I don’t rewrite it for him. He has to work through it. That’s the skill he’ll need no matter what, because real life is just one long series of fixing small mistakes. AI can’t do that part for you.
The third is how to ask for exactly what you want. Leo’s first question to AI for the bee project was “tell me about bees.” That gave him a generic 10-paragraph essay that he couldn’t use for anything. It took him 10 tries to get a useful answer: “what are the top three causes of native bee decline in western Oregon, and list 5 easy things homeowners can do to help that don’t require a big garden?” AI only gives you what you ask for. If you can’t name what you actually need, you get useless, generic output. I practice this with him all the time, not just for AI. When he comes to me and says “I don’t get my math homework,” I don’t just walk him through the whole problem. I ask him, “what part don’t you get exactly?” It’s a small habit, but it teaches him to narrow down what he needs instead of asking for a full pre-made answer.
Last week, Leo finished his bee project, and he added the part about counting bees at the park that AI didn’t have. It got a B+, and he taped it to the fridge next to a photo of his finished lopsided birdhouse hanging in our backyard. I still don’t know what the future holds, and I don’t have all the answers. Some days Leo still argues that he shouldn’t have to do math problems by hand when AI can solve them. I just tell him that the point isn’t to get the answer—it’s to know when the answer AI gave you is wrong. So far, that’s working well enough. We’re still figuring it out, one mistake and one cross-checked fact at a time.
Please indicate: Thinking In Educating » What Should I Tell My Child to Learn in the Age of AI