THE MODEL
Two hours of mastery. Four hours of being human.
A plan for an AI-native Australian primary school, written by a NSW teacher and parent and published openly so it can be argued with and improved.
PrimarAI is a design, not yet a campus.
Everything on this page describes the school as it is planned: where the idea comes from, how the morning and the afternoon work, how students are assessed, and what every child will be able to do by the time they leave Year 6.
The full Blueprint is being published fortnight by fortnight through 2026 under a Creative Commons licence, free to read, free to argue with and free to copy.
HOURS OF MASTERY
2
9am
One school day, 9:00 to 3:30, with recess and lunch in between.
10
HOURS OF BEING HUMAN
4
11
12
1pm
3:30pm
Anything an AI can teach a child, the AI should teach. Anything an AI cannot, a teacher should, and we should make far more time for that than we currently do.
The principle of the school, in one sentence.
AI for the practice
Humans for the meaning
A tool, not a friend
An AI tutor is better than any human at one specific job. It can sit with a child on the same fraction problem for forty-five minutes without sighing, it pitches every task at that child's actual edge rather than the class median, and it never judges.
When you ask children at AI-native schools what they like about their tutor, the answer is almost always that it does not judge them. The morning is built around that.
Reading a child's face. Holding a standard when they want to give up. Noticing that a quiet kid has gone quieter this week. Making a class feel like a tribe.
Only a person in the room can do these things, and they have been squeezed out of teaching for decades by the workload of content delivery and the admin around it. The afternoon gives teachers the time to do the work they came into the profession for.
There are no AI companions here and no chat-as-friend arrangements. A three-year Harvard Business Review study found that the most common use of consumer AI, two years running, is therapy and companionship, and an always-on, always-affirming AI is not what a child needs while they are still working out who they are. Children do that by bumping up against other humans who do not always agree with them. The AI stays a tool. The relationships are with people.
ORIGINS
Where the idea comes from
Three things came together in the second half of 2025 that pushed this from a private project into something that ought to be in public. Each of them solved a different part of the problem, and none of them solved the whole of it for an Australian primary school.
Alpha School, Texas
The shape of the day
An American micro-school running a two-hour, mastery-based academic morning and a four-hour life-skills programme, now growing past twenty-five campuses.
It costs about $40,000 a year in the United States, its data is messier than its marketing suggests, and its culture is more entrepreneurial than most Australian parents would want for their kids.
It has also done the hardest and bravest thing in education, which is to demonstrate that a different shape for the school day is possible, and viable. Its measure of success says everything about its ambition: students should be disappointed when the holidays arrive.
What PrimarAI takes: the 2+4 structure and the mastery pedagogy, and very little else.
What is borrowed, and what is left behind
Feature
Alpha School
Estonia AI Leap
PrimarAI
Two-hour mastery morning
Four-hour afternoon
Life skills, entrepreneurship, TEDx-style talks
No
The five Cs, anchored in NSW: Country, harbour, bush, civic life, older Australians
AI style
Where the AI lives
US consumer platforms
National infrastructure
Australian-built, hosted in NSW, no training on student data
Who meets it first
Age group
Kindergarten to Year 12
Secondary
Kindergarten to Year 6
How it is shared
Yes
Positive, encouraging coach
Students
Private, fee-paying
AI Leap, Estonia
The pedagogy
In September 2025 Estonia launched a national three-year programme that put purpose-built educational AI in front of 20,000 secondary students and 3,000 teachers.
The rule at its centre is different from Alpha's tirelessly positive coach. The AI must ask Socratic questions rather than give answers. It is the most serious national-scale AI education project anywhere in the world, it rolled out to teachers before students, and it runs on infrastructure the country controls.
As its chief executive Ivo Visak puts it, they are building the plane while it is flying.
What PrimarAI takes: the Socratic rule, the teacher-first sequencing, and the insistence on purpose-built, sovereign infrastructure.
No, AI runs inside normal lessons
Socratic, questions not answers
Teachers, then students
Government programme
The Castlereagh Statement, University of Sydney
The Australian frame
Published in October 2025 and signed by university leaders, school principals, public servants and Indigenous educators, the Castlereagh Statement sets out six principles and three horizons for Australian education in the age of AI.
Three of the six speak directly to primary school: redefine what a future-ready Australian is, build curriculum around capability rather than content, and make technology serve pedagogy and trust.
Its far horizon imagines restructuring schooling around student interest, in closer collaboration with families and community. It then leaves the primary school version as an exercise for the reader.
What PrimarAI takes: the principles and the horizons. PrimarAI is an attempt to do that exercise in public, showing the working as it goes.
Yes
Socratic, questions not answers
Teachers, then students. No child meets the system before their coach has spent a term inside it
Published openly under Creative Commons, designed for NESA registration with a scholarship pathway from day one
THE STORY SO FAR
How we got here
This did not start as a plan for a school. It started with a question in a classroom, and it has been pulled along since by what other people have been building.
One hot Wednesday
2025
September 2025
October 2025
June 2026
Next
Can you be mean to a chatbot?
The Castlereagh Statement
Late 2025
A weekly ethics class asks a Year 6 group whether you can be mean to a chatbot. You could hear a pin drop, and then everyone was talking over each other. Nobody else was asking them anything like it.
PrimarAI begins
One of the first Australian ventures built specifically to deliver AI incursions in primary classrooms starts running sessions in Sydney schools.
Estonia launches AI Leap
Purpose-built, Socratic AI in front of 20,000 students and 3,000 teachers, teacher-first, on national infrastructure.
The University of Sydney publishes six principles and three horizons for Australian education in the age of AI, with a primary-school-shaped hole in the middle.
Alpha School goes global
The Texan 2+4 micro-school passes twenty-five campuses and puts a different shape of school day into the world conversation.
The plan is written down
The PrimarAI School: A Plan for an AI-Native Australian Primary School is published in full, with the Blueprint following fortnight by fortnight.
The first afternoon
One principal, one NSW primary school, one term of the four-hour afternoon with a real class. That is the ask.
WHY NOW
Designed for the world the six-year-olds will work in
The school in this plan is not designed for the world we are in. It is designed for the world the children starting Kindergarten this year will be working in. That world is being built right now, and three features of it shape the design.
Adapting will not cut it
Castlereagh talks about adapting the curriculum to fit AI in.
In corporate life, the people who treat AI as something to adapt to are being overtaken by the people who treat it as a reason to redesign the business, and the same will be true of schools. Only the ones that change the fundamentals will end up making sense.
Estonia is the early evidence that whole-system rethinking can happen at national scale without losing the things worth keeping.
AGI is not a robot
The METR research organisation has been tracking how long a task an AI agent can complete on its own, and that duration has been doubling roughly every seven months. First we controlled the prompt. Then agents did the job and we monitored them. Now agents monitor other agents. Each step moves the work further from the human, and managing agents will turn out to be a phase, like prompt engineering was.
What is coming will be closer to a colleague than a tool, and the skills that matter are the ones that make a person worth having alongside it.
In glasses and walls and bracelets and cars
Chat-based AI is transitional too. The next interface, already in research preview, works continuously: it watches, listens, interrupts and speaks up when the world around you changes.
Chat at least makes a child articulate a question and evaluate the answer, which is a useful friction. Ambient AI lowers that floor. A school that only prepares children for screen-based AI is preparing them for the past, the way a 1995 computing class taught DOS commands at the dawn of Windows 95.
The capabilities that matter most
For sixty years schooling has run on the idea that education exists to prepare children for work, and for a long while that bargain was honest, because a child in 1975 could see the shape of the job at the far end of the tunnel.
The OECD's Education for Human Flourishing framework now concedes that the human capital model prepares children for yesterday's jobs and calls instead for the human agency that machines do not have. If the AI-saturated world is the one these children will live in, these are the capabilities that matter most in primary, and none of them is what the system currently measures.
Taste and judgement
Knowing what is good and what is worth doing.
Agency
Setting your own direction without being told.
Embodied capability
Physical skill, manual craft, presence in a room.
Relational depth
Really being with other humans, including difficult ones.
Ethical reasoning under ambiguity
Deciding when AI should be in the loop and when it should not.
Asking good questions
Often worth more than producing answers.
Self-knowledge
Knowing what you want and what you will defend.
Synthesis
Pulling threads together across the domains AI handles separately.
9:00 to 11:00
The morning: two hours at each child's edge
The NSW K to 10 English syllabus and the K to 10 Mathematics syllabus are the spine of the academic morning. An AI tutor, Australian-built and hosted on Australian infrastructure, knows where each child sits against every Stage 1, 2 and 3 outcome, and generates a lesson pitched at that child's actual edge rather than the class median.
A NESA-accredited teacher sits alongside as a coach, pulling children out for conversations, holding standards, and watching for the kid who is gaming the system.
Two hours a day, every school day, is more focused, personally pitched academic work than most children get in a full week of whole-class teaching.
A Year 4 morning, step by step
1
Warm-up
5 MINUTES
Something the child already knows, partly so the engine can calibrate where they are today.
5
Second cycle, then done
BY 11:00
After two cycles the morning is finished. No homework follows it.
Mastery, not marks
There are no percentages and there is no 67 per cent. A skill is either mastered or in progress, and the child keeps working on it until it is theirs for life. No child moves on until they have actually got it, and no child is held back because the rest of the class has not caught up.
Every Stage outcome is tracked individually, so parents always know exactly which capabilities have been secured and which come next. Children work within their NSW Stage group, which keeps the cohort together socially, but the academic progression happens at each child's own pace.
Mastery learning is the most evidence-supported approach in education research, and the school uses it for that reason.
AI for the practice, humans for the meaning. English K to 6 as the worked example
English is where the AI question bites hardest, because a chatbot can produce a passable persuasive paragraph in five seconds. Under the current NSW syllabus, children develop oral language, phonics, reading fluency, comprehension, vocabulary, spelling, handwriting, writing and creative composition. Here is how the model splits that list.
The morning, with the AI tutor
Phonics and grapheme practice
Decoding
Spelling
Vocabulary acquisition
Sentence-level grammar
Reading comprehension at the right level for the child
Oral fluency prompts
The same pattern, AI for the practice and humans for the meaning, is applied to Maths, Science and Technology, HSIE, PDHPE and Creative Arts in the Blueprint.
2
Working block
15 MINUTES
The next concept at the child's edge. The tutor uses Socratic prompts in the Estonian style, not direct answers.
3
Stuck for two minutes?
AS NEEDED
The system flags it on the coach's screen. The coach walks over and talks it through in person.
What the teacher does
Nobody teaches from the front. The teacher moves quietly between students, runs small-group conferencing for children who have just finished a learning unit, and pulls individuals aside for a five-minute check-in on the piece of writing they have been working on at home.
The teacher might notice that a Year 6 child has been quietly bluffing her way through long division since Year 4 and put her on a fast backfill. The teacher is the one who sits beside her while she does it. This is not a reduction of teaching. It is a rescue of teaching from the machinery of the timetable.
The afternoon, with a teacher
Writing as a deliberate, iterative, human-led practice
Composition with real purpose, for real audiences
Discussion of texts in a circle, with classmates and an adult who has read them
Speaking and listening with the actual people in the room
The teacher reading the child's writing line by line and talking them through the redraft
AI used for feedback drafts and idea generation only, never for the conversation about what the writing is trying to do
4
Break
EVERY 25 MINUTES
Off the screen, briefly, then back into the next block.
TRUST
Privacy is central, not bolted on later
A school that asks parents to send their children into an AI-rich environment owes them this much by default. The alternative, which is American consumer-grade AI dropped into Australian primary classrooms, harvesting data on six-year-olds and training the next model on their work, is unacceptable in a way that does not need to be argued. The cost of purpose-built Australian infrastructure is real, and it is unavoidable. These four commitments are written into the school's charter, not just its privacy policy.
The data stays in NSW
Hosted on Australian infrastructure, resident in New South Wales, compliant with the Australian Privacy Principles.
Teacher-first, not student-first
Trust has to be sequenced, not assumed. When Harvard handed its students free AI licences, uptake stalled, because students suspected the tool was a trap to catch them cheating. That is part of why this model rolls out to teachers before children. No child meets the system before their coach has spent a term inside it, and the teachers' programme includes training, school-based learning communities and subject-specific materials developed with subject teacher associations.
No model is trained on student input
Nothing a child types, says or makes is used to train any AI model. Ever.
Parents can see everything
Every piece of data the system holds on a child, in plain language, available to the parent on request, with deletion on request at any time.
Tools, not feeds
The screens in this school are tools rather than feeds. The tutor runs in a closed system with no notifications, no social media, no screen-streaming of children and no corporate spyware sitting on top. A working session, then off. Two hours of focused academic work is less screen time than a NSW Year 4 child currently gets in an average school day, and far less than the bus home. The afternoon is screen-free unless a child is using AI as a tool for a project, under supervision and as part of a deliberate verification practice.
An independent audit, every year
An annual independent privacy audit, published openly, so nobody has to take the school's word for it.
11:30 TO 3:30
The afternoon: four hours of being human
The afternoon is the part of the day where AI mostly stays out of the way. It is not soft padding to balance the morning. It is the part of the day that gives the rest of the design its meaning. Five capabilities, the five Cs, are each taught explicitly, year on year, with measurable progression and external demonstration, in something closer to half the school day than a ninety-minute window once a week. Everyone in 2026 is being sold frictionless. The afternoon is friction by design: real audiences, real budgets, real teammates who disagree, and real consequences when something fails. Children become themselves by bumping up against the world.
Frequently asked questions
Grit, year by year
2km
Year 1
The half marathon at the end of Year 6 is a graduation requirement.
Alpha runs a similar structure in the United States around teamwork, storytelling, entrepreneurship, relationship building and grit. The structural idea is the same. The cultural setting is not. By the end of Year 6, these children do not deliver TEDx talks or secure $10,000 of start-up funding. They speak at the local council meeting, run a market garden, swim a kilometre in the harbour, sit a yarning circle with an Elder, and acquit a $500 community project budget. The form is borrowed. The substance is decisively local.
5km
Year 3
10km
Year 5
21.1km
Year 6
The public speaking ladder
Kindergarten. Tells a story to the whole school.
Year 3. Three-minute talk at the local cafe, to adults who did not have to come.
Year 4. Fortnightly public podcast.
Year 5. Letters to council that get answered. Submissions to government inquiries.
Year 6. TEDx-style talk to a public audience including local media.
Graduation. A ten-minute talk to an unknown adult audience, without notes.
Structure
The structures that hold it together
Most of what makes the day feel different is not the technology. It is a handful of organisational decisions that the technology makes affordable.
Tribes, not classes
Multi-age home groups of about a dozen children spanning the primary years, with a coach who stays with the same tribe from the day they start to the day they leave. The tribe opens the day with a morning circle and closes it with tribe time. It is one of the quiet structural things that makes the school feel different from any other.
No homework
Two hours of academic work in the morning is enough. The afternoon work, which is essentially the project work most schools would set as homework, is done with other people, in the world, where it belongs.
Mastery within Stages, not year groups
Age-grade progression is an administrative convenience, not a pedagogical principle. There is no evidence that chronological age predicts what a child is ready to learn next, and overwhelming evidence that mastery of prerequisites does. Children work within their NSW Stage group and progress at their own edge. NESA's own Stage 1, 2 and 3 structure is already half-way there.
Out of the building, often
Once a fortnight the whole school goes off-site for the day. Once a term there is an overnight camp. Once a year Stage 3 does an extended wilderness immersion in the Blue Mountains and Stage 2 does a week-long exchange with a Western Sydney primary, with families hosting in both directions.
Coaches, not deliverers
Every coach is a NESA-accredited teacher, paid significantly more than teachers currently are, with smaller groups and higher trust placed in their professional judgement. The economics of fewer staff at higher pay work when the rest of the structure is doing what it should. These are the teachers currently leaving the profession because the job has become photocopier-jockey for a curriculum AI has already disrupted underneath them.
Three coordinated programmes
A teachers' programme: training, school-based learning communities, and materials developed with subject teacher associations. A student programme: partnerships with student councils, debating societies, Aboriginal land councils, surf life saving clubs and local theatre companies. A technology programme: a purpose-built Australian learning application on Australian infrastructure, with paid access to advanced AI tools for every teacher in the network.
What is kept from the old model, and what is let go
Kept, undiminished
- The expectation that every child, regardless of background, reaches foundational literacy and numeracy.
- The discipline of a shared daily rhythm: arriving on time, finishing what you started.
- The civic role of school as a place where children meet other children unlike themselves.
- The professional expectation that the adult in the room is qualified, accredited and accountable.
Let go
- Whole-class content delivery.
- Lockstep age-grade progression.
- Social promotion regardless of mastery.
- The assumption that children should be assessed by sitting still and writing things down.
What carries forward, undiminished, is the seriousness of the project.
Six layers, not one grade
Assessment
A single number has never told a parent who their child is becoming. What is needed instead is a layered picture that tells parents, teachers and the child what they are becoming, what they can do, and where they are heading next. Six layers, each designed to be honest, useful and difficult to fake.
6
The graduation sign-off.
At the end of Year 6, three trusted adults publicly say the child is ready: their coach, a community Elder, and an external practitioner.
5
The child as a whole person.
Twice a year, a written portrait from the coach, and a report on curiosity, kindness, honesty, perseverance, courage and attention.
4
Knowing yourself as a learner.
From Year 2, children assess their own work, and the quality of that self-assessment is the most important measure of all.
3
The portfolio of a growing person.
A body of work across the primary years, curated by the child, reviewed in conferencing.
2
Real performance, in front of real people.
The assessment is the doing itself: the cafe talk, the yarning circle, the fundraiser.
1
Mastery, not marks.
Every NSW K to 6 outcome is mastered or in progress. No percentages, no 67 per cent.
Together, these six layers produce something traditional assessment cannot, which is an accurate, honest, multi-dimensional picture of a real child. The mastery system tells us what they have learned. The performances tell us what they can do in the real world. The portfolio tells us how far they have come. The self-assessment tells us how well they know themselves. The narrative tells us who they are. The sign-off tells us, in the voice of trusted adults, that they are ready.
The children still sit NAPLAN, and they perform well at it, because mastery-based morning learning develops the underlying skills the test is trying to measure. NAPLAN is one small data point in a much fuller picture, and the school reports enough comparable data that any secondary school can confidently admit one of its graduates.
Graduation
A child does not graduate without them
Every Year 6 graduate of this school can demonstrably do the following fifteen things. These are not aspirational outcomes. They are graduation requirements, signed off in the final term by the child's primary coach, an external practitioner and a community Elder. A parent reading the list knows what they are buying into. A principal reading it knows what they would be running. A child reading it knows what is expected of them and, more importantly, why most of it is worth doing.
1
Read fluently, write a 2,000-word researched piece, and edit their own work.
4
Lead a yarning circle on a topic of their choosing.
7
Cook a meal for thirty people from a planned menu and within a real budget.
10
Make at least one substantial physical object: a piece of furniture, a working device, a sewn or woven item.
13
Hold a sustained relationship with at least one older Australian over multiple years.
Fifteen things. Concrete, demonstrable, and judged externally.
2
Operate confidently across the full NSW K to 6 Mathematics syllabus, with mastery of every Stage outcome.
5
Swim one kilometre in open water and handle themselves around a small boat.
8
Plan, run and acquit a community project with a budget of at least $500.
11
Direct an AI system to do useful work, verify its output, and explain when not to trust it.
14
Write a letter to a politician that elicits a real reply.
3
Deliver a ten-minute public talk to an unknown adult audience without notes.
6
Complete a half marathon or an equivalent endurance event.
9
Play one musical instrument at an intermediate level and perform publicly.
12
Speak basic Bidjigal language and identify ten native plants on local Country.
15
Articulate, in writing, who they are becoming as a learner.
If you only have a minute
What makes this school different from every other primary school in NSW
Two hours of academic work each morning, instead of six.
Country-based learning led by Aboriginal educators, as part of the regular calendar.
Mastery progression rather than age-grade. No child moves on until they have it, and no child is held back because the class has not caught up.
Wilderness immersion that builds from one overnight in Year 1 to a multi-day trip in Year 6.
AI used as a Socratic coach rather than an answer engine. Australian-built, NSW-hosted, privacy-first.
A sustained relationship with one local senior centre, fortnightly, over multiple years.
No AI companions or chat-as-friend arrangements. AI is a tool, not a relationship.
No homework. The afternoon work, done in the world, is the homework.
Multi-age tribes, with the same coach staying with the same children across their whole primary journey.
Fifteen demonstrable graduation requirements, all of which have to be met.
Four hours of real-world capability every afternoon, across the five Cs.
Six layers of assessment in place of a single number.
Public performance baked into the year, with every child speaking to real audiences from Kindergarten.
Teachers paid significantly more, working with smaller groups, trusted with their professional judgement.
Real microbusinesses run by the children, with real money and real customers.
The Blueprint published openly under a Creative Commons licence, so any school that wants to adopt the model can.
THE HARD QUESTIONS
Objections, answered
These are the questions that come up first, every time, from teachers, principals and parents. Here are the straight answers.
Join in
Help me build the next one
I am one teacher in the eastern suburbs of Sydney with a kid in primary school and a calculator. I cannot build this on my own, and there is no point pretending otherwise. Nobody has ten years of experience with AI in primary classrooms, which is not a reason to wait. It is a reason to collaborate, share and build in public, so that the people who are actually doing the work can find each other while there is still time to do it well together.
If you are a teacher who is tired, a principal who is curious, a parent who is worried, a board chair with a half-empty enrolment, an investor who is interested in what comes next, a Castlereagh signatory who quietly wonders what any of this means for an eight-year-old, or just somebody with a strong opinion and a reason to engage, I want you in the conversation. The design only really gets interesting when other people start arguing with it.
The model itself is wrong for the era. Help me build the next one. With teachers, not around them. With children, who, believe me, already know.
Dane Smale, Sydney, 2026
Three ways in
The plan is licensed under Creative Commons Attribution 4.0. You are free to share and adapt it for any purpose, including commercial, as long as you credit the author and link to the licence. Cite as: Smale, D. (2026). A practical plan for an AI-native Australian primary school. PrimarAI.