Every school I visit has two kinds of teachers. The first kind have used AI for more than twenty hours. They don't talk about it much — they've moved past the novelty — but their planning is faster, their differentiation is sharper, and their students are being asked harder questions. The second kind have used AI for less than an hour. They're not behind because they're unwilling. They're behind because no one has given them a reason to spend those twenty hours.
That's the gap. And unlike the tech gaps of previous decades, this one compounds weekly.
The case against "wait and see"
The argument for waiting is almost always a version of "the tools aren't mature enough yet." I understand it. It used to be my argument too, as recently as the spring of 2023. What changed my mind was a sequence of visits to schools who had committed early. Their teachers weren't coping — they were enjoying it. Their students weren't resentful — they were being stretched. And the teachers who had been skeptical six months earlier were, to a person, glad they'd been pushed.
Waiting assumes the cost of the delay is zero. But every week a school defers AI fluency is a week its most curious teachers spend building that fluency alone, in isolation, often without their colleagues knowing. You end up with a pocket of advanced practice and a much larger pocket of untouched practice, and the gap between those pockets is bigger than any external gap between your school and someone else's.
What faculty-wide fluency actually looks like
It's not a workshop. Workshops end.
Faculty-wide fluency is what you get when:
- Every teacher on your faculty has planned at least three full lessons with AI assistance.
- Every department has had at least one conversation about which of their outcomes are AI-proof, which are AI-vulnerable, and which could be AI-amplified.
- Leadership has drafted, published, and iterated on at least one version of the school's AI policy — not the final one, just the first one.
- Parents have been given a plain-English explanation of what AI is being used for in their children's education, and what it is not being used for.
None of those are ambitious on their own. Together, they produce a school where the default answer to "can we use AI here?" is a thoughtful "yes, and here's how" rather than a nervous "we're still figuring it out."
The economics of the hours
A typical faculty has somewhere between forty and two hundred teachers. If each of them spends twenty hours getting meaningfully fluent with AI — call it an hour a week across a term — you're talking about a thousand to four thousand hours of teacher time. That sounds like a lot until you realise it's about as much as a school spends on a single curriculum-review cycle.
The difference is what you get back. A curriculum review produces a new document. Faculty-wide AI fluency produces a school whose every lesson, from next week on, is slightly better than it would have been otherwise. The compounding matters.
What happens if you don't
Schools that delay this decision past 2026 will, in my experience, find themselves having the same conversation three years later, with one critical difference: by then, parents will have noticed. Today's early-adopter parents are already comparing schools on this. They're asking specific questions at admissions interviews. And the schools that can answer those questions in concrete terms — "here's what our Grade 9 science sequence looks like now, here's what it looked like before, here's what our teachers say about the change" — are the ones they're choosing.
That's not a prediction. It's already happening, and it's visible at any admissions event where an informed parent is in the room.
The decision isn't whether your school will do this. It's whether your school will do it before, during, or after your competitors.