Part One of Four. Schools are being asked to decide how children meet artificial intelligence. The honest answer is not a yes or a no. It is a question of timing — and that is what this series, and the work behind it, is built around.
There is a generation of children in school right now, and decisions about how they meet artificial intelligence are being made for them this year. They are made in classrooms, in policies, and in the quiet defaults of what teachers do and do not allow. Most are being made without a clear way to think about them.
The question schools are really asking
When a head of school asks about AI, the question underneath is almost never “is this good or bad.” It is when. When is a child ready to use this? When is it too soon? What belongs in Grade 7, and what should wait until Grade 11?
Most current frameworks offer competence inventories: careful, useful lists of what students should be able to do with AI. But a list of competencies describes a destination. It does not tell a school the order of the journey, or when each step is safe to take. That is a question about how children develop, and a list is not built to answer it.
A different starting point
Our work begins from the premise that the way schools introduce AI should follow the way children actually develop. A child’s mind is built in a sequence. Some capacities consolidate early and become the ground everything later stands on. Others — judgment, the ability to weigh evidence, to monitor one’s own thinking — mature slowly, well into late adolescence. A framework for AI in schools can honour that sequence. When it does, it can finally answer the question of when.
The reasoning, and the honesty it requires
There is a hard fact at the centre of this. The evidence that would tell us, with certainty, how AI use shapes a developing mind will not exist for fifteen or twenty years. A childhood is the unit of measurement, and we are inside one.
That fact is often used as a reason to wait. We think that is a mistake. Waiting is not a neutral act. It is itself a decision, taken without evidence, and its cost falls on the children who are in school while the waiting goes on. The responsible path is not to wait for proof, and not to rush ahead of it, but to reason carefully from what is genuinely known: how children develop, and how earlier technologies have changed the cognitive work of learning.
What this series does
Over four articles, we set out that reasoning in the open. This first piece establishes the claim: the real question is when. The second sets out the mechanism that makes timing matter — what cognitive offloading does inside a mind that is still forming. The third follows the quieter substitution: AI standing in not just for a child’s work, but for their company. The fourth describes the response — staging, literacy before fluency, and refusal as a skill worth teaching.
Behind the series sits one body of work. The Position Paper carries the full argument. The framework built on it — AI by Design — is the developmental architecture: a structure that stages AI across grades 6 to 12 by how students mature, rather than by a list of competencies or a ladder of task difficulty. The curriculum is what that framework looks like in the classroom, staged year by year. Each will be published in turn.
Taken together, the philosophy, the paper, the framework and the curriculum share one conviction: that AI should extend a young person’s capability, not replace the effort that forms their mind.
Where to begin
If the question of when a child is ready for AI is a live one in your school, the Position Paper is where the thinking starts. It sets out the argument, the evidence behind it, and what it commits the framework to.
Read the Position Paper — The Formed Mind: AI by Design. [Position Paper Here]
Next in the series: What a Mind Never Does, It Never Builds — cognitive offloading and the forming mind.
The EdJournal · Education. Intelligence.