Using AI to Write a Paper About AI (But Not the Way You Think)
My high schooler is writing a paper on AI in education. The good, the bad, the risks. We are sitting at the kitchen table working through it the old way: thesis statement, introductory paragraph, build the argument, land the conclusion. His textbook is open. His notes are spread out. Next to him, my laptop is open to Claude. Neither of us pretends that isn’t strange.
AI is woven into how I think, draft, revise, and do my work. I write about it. I build with it. I advocate for its thoughtful use in social impact and evaluation. And here I am telling my teenager to put the tool down and do it by hand. I haven’t fully resolved that contradiction. I’m not sure I need to. But I owe him honesty about it, and honesty starts with admitting that I’m figuring this out too.
Here is what I do know. AI can genuinely help kids learn. It can explain a concept five different ways until one clicks. It can help them organize messy thoughts. It can give feedback on a draft at midnight when no teacher or parent is available. I have seen it work. But it can also let them check out completely. Offload their thinking. Get a polished paragraph back without ever having struggled with an idea long enough to actually understand it. The struggle is where the learning lives. Skip it and you get a good-looking paper from someone who didn’t learn anything writing it. AI is not one thing. The question is whether the kid is still thinking or whether the kid has stopped thinking and doesn’t realize it yet.
So we are doing the process. Devising a thesis statement that actually says something. Writing an introduction that sets up the argument instead of just restating the prompt. Building the body with evidence and reasoning. Writing a conclusion that doesn’t start with “In conclusion.” It is not glamorous. It is the same process my professors taught me, and it works. He is researching the risks of AI in classrooms while I sit next to him using AI to run my practice. We talk about that. It is part of the education.
I am not banning anything. “Do as I say, not as I do” doesn’t hold up when your kids watch you use the thing every day and know exactly what you do with it. What I am trying instead is a sequence: learn the process first, then use the tool. You have to know what good writing looks like before you can recognize whether AI gave you good writing or just confident-sounding writing. You have to have built an argument yourself before you can evaluate whether the machine built one that actually holds together. You cannot manage the work if you have never done the work.
I don’t have this figured out. I am parenting through a transition I cannot see the other side of, and anyone who tells you they have the answers is selling something. But I know that sitting at the kitchen table working through a thesis statement together is not wasted time, even if the world my kid graduates into looks nothing like the one I learned to write in. The process is the point. Not because the old way is sacred, but because you cannot use a powerful tool well if you never learned what the tool is doing for you.
Anthralytic is a strategy and evaluation studio that helps mission-driven teams clarify and amplify their impact using human expertise, data, and AI.
