Why I Use AI Every Day—and Why Schools must Teach It

I found ChatGPT in November of 2022, and I’ve used AI nearly every day since. At first, I used it to understand the technology well enough to help schools implement it effectively. Now it’s simply part of how I work. I use it to process large data sets, chase down research, bounce ideas around for posts — including this one — and make my working life more efficient so the rest of my life gets more of me.

And in every one of those use cases, I know exactly what I’m doing. I know what the tool collects about me and about the work I feed into it. I know it isn’t always correct, and that on plenty of tasks the most efficient move is to put it down and think without it. I know it can be annoyingly agreeable when I float an idea, so I’ve learned to push back on it the way I’d push back on a colleague who agrees too quickly.

It’s not perfect. Neither am I. But I understand my why for using it, I understand how it works, and I understand what it costs me. Why, how, and at what cost — that’s the whole ballgame, and it took me three years of daily, intentional use to build that judgment.

So when I see the growing call for AI-free schools — the same energy now pushing to ban all screens — I sit with one question: who is going to prepare our students the way I prepared myself?

Because that’s what we’re actually missing when we ban something. We defer the work. We defer the work of preparing students to use technology effectively and appropriately, and we hand that work to nobody — which means students do it alone, at home, unguided, and badly. And here’s what makes it worse: the loudest voices calling for these bans use these tools daily. They’ll cherry-pick storylines and conflate them with what’s happening in their schools. Jonathan Haidt does this with his call for cell phone-free schools — which, for the record, I completely agree with — and then stretches it into screen-free schools, which is a blatant disservice to our students. The phone in the pocket and the tool in the lesson are not the same problem, and pretending they are is how good policy goes bad.

So, with all of that said: how do we bring AI into our schools effectively?

I think we look no further than the work we already have to do — recalibrating the 1:1 programs that are under fire across the country. The discipline is identical. We got 1:1 wrong by leading with procurement instead of purpose, and we’re paying for it now with audits, backlash, and bans. AI gives us the chance to run the play correctly the first time. Here’s the rough shape of it.

Start with the why, before any license. Your school has to be able to say, out loud and in writing, what it believes AI is for — a thinking partner or an answer machine, where assistance ends and authorship begins, what original work means in your building now. If your leadership team can’t answer that in plain language, no acceptable-use policy will answer it for you.

Write policy as a living document, not a permission slip and not a ban in legal language. Who gets access, under what conditions, with what guardrails — reviewed on a cycle, because this technology will not sit still for your three-year plan.

Make your faculty fluent before your students log in. Teachers need access, time, and permission to be novices first. Give them the same journey I gave myself: daily use, real tasks, honest encounters with what the tool gets wrong and when to put it down. A teacher who has lived that can teach judgment. A teacher who got a login the same week as their students can only teach compliance.

Then, and only then, students — inside a fluency progression, not a free-for-all. Prompt well, verify everything, own what carries your name, and know when to walk away from the tool. Built developmentally, starting with concepts before devices, the same way we should have built 1:1.

Notice that every phase is the answer to the ban crowd’s legitimate concern. They’re right that unguided AI use is a problem. The ban just guarantees the use stays unguided. The sequence is what guided looks like.

I didn’t become a fluent AI user by being kept away from it. I became one through three years of intentional, daily, critical use — with my why defined and my eyes open. Our students deserve the same path, with something I didn’t have: adults who planned it.

This week’s resource for paid subscribers is the deep dive on exactly that: the AI implementation sequence, built around making your faculty and your students fluent — the phases, the exit criteria for each, and what fluency looks like at every level of your building. The argument is free. The sequence is in Think Forward In Practice. Coming later today.