So that is where we begin, not with the technology, but with the child. Every decision about AI at The Civic School starts from the developmental science of childhood and works outward. This page lays out our approach.
The elementary years are when the brain builds attention, language, executive function, and the felt sense of one’s own mind. These capacities have developmental windows and we know skipping the building doesn’t get redone later. We build them first and protect them fiercely.
AI’s first job here is to help adults see children more clearly. Whether in planning, in observation, or in communication with you it is never a substitute for a teacher, and it is never a companion for a child.
Every tool must earn its place in a child’s day. AI enters children’s hands on a deliberate, published ladder: co-engaged with a teacher at first, always bounded in time, followed by reflection on what the child did and what the tool did.
Our first graders will be adults in 2040. The capacities that will matter then - judgment, voice, persistence, presence, self-knowledge - must be built first, so that AI fluency sits on top of a fully formed person rather than substituting for one.
The long-term effects of growing up alongside AI are not yet established as the technology is too new. But the developmental science of our development is not new, and it is unambiguous about what young brains need to build.
A child’s capacity for deep, sustained attention grows through years of practice and the gap between attention that is coerced and attention that is chosen is where schools succeed or fail. We design whole afternoons around work children choose and stay with, because the brain only builds what it practices.
Reading is not natural or neutral. When a child learns to read deeply, they build the substrate for sustained attention, inference, and the patience to think hard thoughts. At The Civic School children build that circuit early on.
Critical thinking is not a skill floating free of content and children reason well when they understand something with depth. Looking something up is not the same as knowing it, and no search or summary can substitute for a mind stocked with the world. That is why our days are devoted to knowledge worth having.
Planning, deciding, holding a goal, resisting the easy pull - strong executive function is among the strongest childhood predictors of adult flourishing, and it is built through meaningful, goal-directed work with adult scaffolding. Children at The Civic School learn to run their own work before any tool offers to run it for them.
Understanding what another person thinks, feels, and needs is a capacity children construct through years of play, conversation, negotiation, and repair. A chatbot cannot practice it for them and pretending to be a friend is the one thing we never let AI do here.
A dysregulated child is not a thinking child. Emotional regulation isn’t separate from academics; it is the precondition for them. Which is why it’s the first of the seven domains we observe and report through SCOPE™.
We don’t teach reading and math to prepare children for jobs. We teach them because the developing brain becomes fully itself by building these capacities and a child who doesn’t get to build that brain doesn’t get a second chance at it.
Children here don’t stumble into AI; they climb toward it, with a teacher alongside. Here is what that actually looks like, year by year.
Children never use AI alone in the early grades. It enters in narrow, teacher-guided moments: dictating a story aloud and revising it on paper, watching AI attempt to draw what the class observed and critiquing the result, asking AI to summarize a text the class has read (and catching what it got right or wrong). Children always draft their own work first.
By the end of second grade, children can say at their own level what a language model is: a pattern-completer trained on human writing, which does not know things the way they know things.
Scaffolded individual use begins under standing classroom protocols. Any AI-generated fact must be confirmed in two non-AI sources before a child may use it. Any sentence shaped by AI is marked in the draft. Children keep an AI-use journal: what I asked, what I got, what I kept, what I discarded, and why.
The sentence that organizes everything from here forward: “First I think, then maybe I check with AI, then I think again.”
Multi-tool projects span weeks and combine AI with books, primary sources, interviews, and direct observation. Children learn where training data comes from, what a hallucination is and why it happens, and how bias shows up. Children are actively testing it themselves by comparing how AI answers questions about different communities and places.
And the first explicit lessons in what AI cannot do: it cannot remember you, it cannot be your friend, it cannot tell you who you are.
By fifth grade, AI is a known system subject to scrutiny. Children can explain how a language model works, name its failure modes, and weigh its real costs: labor, environmental, intellectual. The year-end exhibition includes an AI-ethics component in which each child takes and defends a position on a real question.
They practice two disciplines with equal seriousness: using AI well, and choosing not to use it.
Every AI interaction in grades 1–2 is co-engaged with a teacher and followed by discernment work: what did the AI do, what did I do, and what is the difference?
We do not use emotional or relational AI products with elementary-age children. Human relationships include friction, disagreement, and conflict so the risk of a child confusing a machine for an easy relationship is real, and we treat it that way.
School AI use happens on shared classroom devices, in shared time, with shared accountability, never solitary in the social sense.
We open AI use deliberately, close it deliberately, and always end by reflecting on what the child did and what the tool did.
AI is never a substitute for a genuine attempt. A child must try, and be able to say what they tried and where they got stuck, before any tool enters the picture.
Our published screen-time budgets are caps, not targets and even at the upper-grade caps we sit well below typical elementary exposure. Screens never displace instruction, recess, or one another. And we are asking families to do the same at home.
While children climb the ladder slowly, our adults use AI fluently and openly. Teachers narrate small pieces of their own AI use to model discernment: “I asked the AI to suggest a song about today’s weather. Here’s what it gave us. What would you change?”
That’s the inversion that makes our approach different from both the AI-tutor schools and the tech-free schools: we are not avoiding this technology, and we are not handing it to seven-year-olds. We are putting it where it belongs: in the hands of professionals and in service of seeing each child.
Teachers use AI for lesson planning, materials, and family communication freeing their time for what matters, focusing on our children’s education.
AI-assisted analysis helps teachers synthesize their daily observations of each child and a teacher reviews and confirms every conclusion before it is recorded. See how SCOPE™ works →
Decision-making and ethical thinking is one of the seven domains we teach and measure because in an AI age, one of the most consequential things a school can teach is how to choose well.
A child leaving The Civic School is not AI-naive; they are AI-fluent the way we want adults to be fluent: as users with judgment, not as dependents.
At a fifth-grader’s level of detail, with accuracy.
Drafting, brainstorming, transcription, and narrow research, always with verification.
Hallucination, bias, the absence of understanding.
When their own thinking is what’s needed — and they can say why.
By asking of any text or image: who, or what, made this?
Presence, voice, persistence, self-knowledge, care, and craft. The capacities AI cannot give them are beginning to take shape.
The long-term effects of growing up alongside AI are not established. The technology is too new, and the children who will have grown up with it are still small. We are building this school in exactly the moment those effects are being formed and we take that responsibility seriously.
So we hold our approach the way we teach children to hold any claim: with evidence and with humility. Our Screen & AI Use Protocol is a published document, reviewed and reissued every August. We track our students’ development carefully and show families the evidence and we listen closely to what you see at home. If the science changes, we will change with it. And what we learn, we share.
Read our Screen & AI Use Protocol (PDF) →