Curriculum + Make · Complete on its own · Grades 4–12
AI skills for school now. Human judgment for the work ahead.
When AI can do more of the work, young people need to do the thinking that matters. Own Mind teaches what AI is doing, where it fails, how to direct and test it, and what must remain human. Every idea is taught through a content-rich lesson, then put to work in something worth making.
The curriculum architectureOne coherent 4–12 journey
What students actually learn
AI knowledge. AI practice. Human judgment.
Students learn what AI is doing, where data and bias enter, why fluent answers can still be wrong, how to direct and test a machine, and where human authority must remain. Choose an area to inspect the substance.
AI practice · Understand
Know what the machine is doing—and what it is not doing.
AI mental modelsData and representation
Why this matters in AI-shaped work
Knowing when machine output can support a decision—and when expertise, evidence or a different method is needed.
- Models generate likely patterns; fluent output is not knowledge or evidence.
- Training data, labels, representation and omission shape what a system can produce.
- Confidence, agreement and polish can all appear when an answer is wrong.
- Distinguish an AI output from a source.
- Name uncertainty and recognize when “cannot tell yet” is the responsible answer.
- Inspect bias, omission and provenance before relying on a result.
The AI curriculum
Students learn the technology, the judgment and the responsibility together.
AI literacy is taught as knowledge, judgment and responsible action. Students build an accurate mental model, see how data and evidence shape an output, learn to give AI bounded work, and remain responsible for the tests, decisions and consequences that follow.
See what the machine is doing.
AI mental models · data and representation · truth and evidence
Give the machine a bounded job.
Briefing and delegation · ownership of thinking and consequential decisions
Put the exact result under pressure.
Testing and revision · creation and craft
Use AI responsibly with other people.
People and inclusion · collaboration and leadership · venture and governance
Make first → Delegate → Verify → Defend.
The student begins the thinking, gives AI only a defined role, tests what returns and explains the final judgment independently. Every project rehearses this durable working method through two or three focal human capabilities.
The Learning Bridge
Learn it in OWN MIND. Use it where subject knowledge matters.
A move is taught first in the AI curriculum. The Learning Bridge then places it inside a subject problem already worth understanding. Disciplinary knowledge, sources and criteria determine what good looks like; the familiar AI method helps students do the thinking without giving it away.
Recall the AI move
The student retrieves a move already learned in the AI curriculum: source checking, bounded delegation, uncertainty, testing, disclosure or another named habit.
Connect it to the subject
The live topic, disciplinary knowledge, sources, misconceptions and success criteria set the intellectual standard.
Make the human judgment
The student forms the interpretation, prediction, method, priority, intent or design decision before any machine contribution.
Use, test and defend
AI performs one declared job. The student tests the result through the discipline, revises the work and explains the final decision with AI closed.
Previously taught AI move+Subject knowledge+Student-owned task=Visible transfer
Which intervention should our town prioritize to reduce unequal heat risk?
The AI curriculum has already taught claims and evidence, representation, bounded delegation and exact testing. Geography now gives those habits disciplinary weight.
- Geography learning
- Urban heat, spatial inequality, intervention evidence and the distribution of benefit and burden.
- AI’s bounded role
- Cluster the supplied evidence by intervention and reach after the student chooses the priority group and criterion.
- Student responsibility
- Set the initial priority, test the evidence, identify a missing perspective, revise the recommendation and defend the trade-off.
- What remains visible
- First view, sources, machine contribution, keep/change decision, revised recommendation and AI-off explanation.
Across the timetable
One set of habits. Different disciplinary demands.
What interpretation can I defend?
- AI learning carried in
- Write the interpretation first; ask AI for critique through one lens; test every suggestion against the text.
- Machine role
- Critique the student’s claim against one declared lens after the first interpretation exists.
- Human judgment in use
- Critical reasoning: test the critique against the text · Metacognition: explain why the interpretation changed or held
- Work students leave behind
- Interpretive claim · selected evidence · revision note · AI-off defense
What result would support—or change—this prediction?
- AI learning carried in
- Predict before assistance; define the evidence test; revise only when the result warrants it.
- Machine role
- Challenge one test condition or help compare the student’s result with the stated prediction.
- Human judgment in use
- Critical reasoning: connect the result to the prediction · Adaptive thinking: revise the explanation when evidence changes
- Work students leave behind
- Prediction · test design · result · revised explanation
What does this source allow me to claim?
- AI learning carried in
- Judge provenance first; corroborate across approved sources; name what the evidence cannot establish.
- Machine role
- Identify one checkable gap across an approved source set after the student’s first provenance judgment.
- Human judgment in use
- Critical reasoning: bound the claim to the sources · Ethical judgment: name the missing perspective and its consequence
- Work students leave behind
- Source judgment · corroboration · bounded claim · limitation
Where does my method work—and where does it break?
- AI learning carried in
- Show the first method; identify the exact point of difficulty; request only the smallest useful help.
- Machine role
- Ask one next-move question at the student’s declared point of difficulty.
- Human judgment in use
- Self-regulation: seek only the smallest useful help · Metacognition: explain where the method broke and how it recovered
- Work students leave behind
- First steps · point of difficulty · completed reasoning · explanation
Who should this decision serve, and on what evidence?
- AI learning carried in
- Choose the priority group and criterion first; inspect representation; test whose evidence or experience is missing.
- Machine role
- Cluster approved evidence after the student has chosen the priority group and criterion.
- Human judgment in use
- Problem-framing: choose the group and criterion · Ethical judgment: weigh benefit, burden and missing voices
- Work students leave behind
- Priority · criterion · evidence-backed recommendation · trade-off
What is my intent, and what work belongs to me?
- AI learning carried in
- Fix the creative intent and source boundary; define any generative role; compare each version with the intent.
- Machine role
- Generate or critique within a student-defined creative brief and authorship boundary.
- Human judgment in use
- Adaptive thinking: compare versions against intent · Ethical judgment: protect authorship and disclose the machine role
- Work students leave behind
- Creative intent · boundary · iterations · selection and rationale
Three views of the curriculum
See the progression. See every skill. See what students do with it.
Follow the age-band end points and ten-strand progression, explore the complete AI learning map, then see how those ideas become useful work in Maker and travel into subjects through the Learning Bridge.
One connected learning journey.
See what students learn about AI, how the seven habits of mind are explicitly taught, and how both travel into subject work through a clear 4–12 progression.
The complete 61-experience source map.
Browse the current Workings learning spine by age and skill family. Every entry names the concept, the misconception it challenges and the piece of evidence the student makes.
Seven families of useful output.
Explore physical, digital, scientific, creative, service and venture work—with a protected student decision, bounded machine job, real test and responsible return.
Inside a deep lesson
A project brief is not a lesson.
This worked grades 6–7 lesson teaches prediction, evidence, provenance and uncertainty before asking students to perform. Open each stage to see the knowledge, misconception, AI skill and human move—not just the activity.
Curriculum + Make · grades 6–7 · worked lessonWhen confidence is not evidence · Source-backed school guide
Current job · Orient
Begin with a genuine information problem.
A new student needs a reliable guide from reception to the library. Two polished AI answers disagree about the route, so confidence alone cannot decide what goes into the guide.
- Content and vocabulary
- Question · intended user · what a reliable guide must establish
- Misconception confronted
- “The more detailed answer is probably the right one.”
- AI knowledge and skill
- Treat an AI answer as a claim to investigate—not a source.
- Human capability in the work
- Problem-framing: define the real need and the decision that remains theirs.
What remains with the workPurpose · user · success criterion
My school-access guideVersion forming
Getting from reception to the library
Two confident answers disagree. What evidence would make this guide reliable?
Question framedPerformance not yet requestedMachine closedTest ahead
Inside a six-session project
A real need becomes tested work a student can stand behind.
The course retrieves earlier knowledge and teaches what students need next. Every student then makes a first version, decides what AI may do, tests the exact result, reworks it and defends the final judgment.
Find + listen
- New learning
- The course retrieves prerequisites, teaches the situation and models how to distinguish a symptom, an assumption and an evidenced need.
- Student action
- Inspect a source, user account or observation; identify the affected person, purpose and constraint.
- Habits in use
- Problem-framing: separate symptom, assumption and evidenced need · Ethical judgment: identify who is affected
- Evidence retained
- Retrieval response · evidence/assumption split · first problem frame
Mine + make v1
- New learning
- The course explains and models the new knowledge, contrasts a non-example, guides a second case and checks understanding before releasing independent work.
- Student action
- Choose a strategy, plan dependencies and make a substantive first version before delegating anything.
- Habits in use
- Self-regulation: sequence dependencies · Metacognition: choose and justify a strategy
- Evidence retained
- Hinge explanation · plan · strategy reason · first version
Bound the machine
- New learning
- The course teaches model limits, data and provenance boundaries, role contracts, prohibited decisions and human approval points.
- Student action
- Decide whether AI belongs and, if so, specify one exact job and its failure check.
- Habits in use
- Critical reasoning: predict a failure · Ethical judgment: reserve consequential decisions
- Evidence retained
- AI/no-AI decision · role contract · data boundary · approval point
Test
- New learning
- The course teaches how criteria become a source, user, behavior, access, bias or edge-case test and how to interpret an incomplete result.
- Student action
- Test the exact version, preserve the result and compare it with the success criteria.
- Habits in use
- Critical reasoning: run an exact test · Adaptive thinking: interpret uncertainty
- Evidence retained
- Test plan · version tested · result · uncertainty or failure note
Critique + rework
- New learning
- The course guides comparison against criteria and models the difference between cosmetic change, consequential revision and a responsible hold or stop.
- Student action
- Explain what the evidence means, then revise, reject, hold or stop for a stated reason.
- Habits in use
- Critical reasoning: connect evidence to decision · Adaptive thinking: revise, hold or stop
- Evidence retained
- Evidence-to-decision note · changed version · reason for keep/change/hold/stop
Return + own
- New learning
- The course supplies an unfamiliar AI-off task, defense questions and the appropriate return, handover or closure protocol.
- Student action
- Complete the novel task independently, disclose the machine role, hand over or close the work, and defend the decisions.
- Habits in use
- Collaboration: return or hand over well · Metacognition: explain decisions without AI
- Evidence retained
- Individual AI-off task · student explanation · disclosure · handover or closure record
KnowledgePrecise concepts, vocabulary, source bodies and curriculum connections
Responsive teachingA hinge question answered before feedback, followed by the right authored next step
AccessThe same ambitious endpoint through adaptations, offline media and an equal no-AI route
Responsible AIAn approved tool, exact job, data limits, human approval, failure check and safe stop
The bridge from curriculum to work
Prompting is a tool skill. Knowing what not to outsource is a human capability.
Own Mind is designed to give students repeated, increasingly demanding practice in comparable responsibilities. It does not promise a career outcome or label a student ready for the future.
01
Frame what matters
Turn a real need into a precise question, audience and definition of success.
Decide what the machine is—and is not—being asked to solve.
Problem-framing, scoping and making sure a team is solving the right problem.
02
Make before asking
Create a prediction, plan, sketch or first version before requesting help.
Give the student’s intent something visible to protect, compare and improve.
Bringing independent expertise and a point of view to automated production.
03
Judge uncertainty
Trace the claim, test the evidence and state what still cannot be concluded.
Treat confident output as material to test—not authority to obey.
Quality judgment, research integrity and decisions made under incomplete evidence.
04
Direct tools deliberately
Choose AI, another tool, a person or no tool, then set a bounded role.
Specify inputs, exclusions, approval points, tests and a stopping condition.
Responsible delegation and coordination across people, tools and AI systems.
05
Take responsibility
Keep, change or reject a contribution, disclose it and explain the decisive work.
Defend the result with the machine closed and name the limits that remain.
Authorship, accountability and owning the consequences of a professional decision.
The curriculum creates structured practice and inspectable work. It does not score capability, infer a student profile or claim that completing a project proves employability.
Retrieval, practice and transfer
Each protected hour adds to the next.
Each lesson retrieves what came before, adds precise new knowledge and moves it into purposeful work. Interstitial sessions bring important ideas back; Learning Bridge tasks give those ideas a new disciplinary context; later checks ask students to select the move for themselves.
Teach the knowledge
Concepts, vocabulary, source material and misconceptions are taught directly, with a worked example beside a meaningful non-example.
Rehearse the move
Students complete a guided second case, compare strong and weak responses and commit to an explanation before feedback and before the scaffold fades.
Make something that matters
Every student contributes an attributable first version, uses evidence to test it and makes a consequential revision or responsible stop.
Use it somewhere new
Knowledge and habits return through retrieval, an independent check with AI closed and a Learning Bridge task in a changed subject or project context.
Explain → model → practice → make → test → revise → retrieve → transfer. Each return asks more of the student: less prompting, a different context and a more consequential decision.
| Age band | Planned annual time | Cadence | Curriculum shape | Recommended placement |
|---|---|---|---|---|
| 4–5Maker Desk | 32–36 planned hours in the Year 6 model | Weekly, 54–50 minutes | Six-session projects with a concrete, physical and offline-first character. | A protected weekly slot from September to May, with optional post-SATs acceleration. |
| 6–7Contributor Studio | 38–40 planned hours per school year | Weekly, 60 minutes | Five six-session project arcs plus protected retrieval and transfer sessions. | Its own KS3 timetable line—not a short carousel. |
| 8–10Venture Studio | Year 10: 38–40 hours · Year 11: 18–20 hours | Weekly, 60 minutes | Six-to-seven-session projects, with doubles for make and test where the timetable offers them. | A Year 10 personal-development or enrichment line, with a lighter Year 11 transfer year. |
| 11–12Venture HQ | 60 planned hours: about 30 supervised + 30 timetabled studio | Weekly supervision + protected studio | Eight-to-ten-session venture arcs with operational handover or responsible closure. | Sixth-form enrichment, with an optional EPQ-paired extension where appropriate. |
All core learning fits inside the school day, with offline media and 10–15% flexibility built into the annual plan. Learning Bridge tasks happen in the relevant subject lesson after students have learned the AI move in OWN MIND.
Knowledge moves into first attempts, bounded delegation, exact testing, rework and ownership.
Earlier knowledge returns, prerequisites reactivate and the next context comes into view.
Previously taught AI habits meet live disciplinary questions, sources and success criteria.
Students retrieve, select and adapt the move after the original prompts have gone.
The 4–12 journey
Catching a mistake becomes leading a careful system.
Catching a visible mistake at 10 becomes tracing provenance, testing serious work and governing a bounded system at 18. Schools can enter at the age-appropriate band; this is progression, not a prerequisite staircase.
The 4–12 AI journey
A course in the human judgment AI can't do for us.
Students learn how AI produces output, where data and bias enter, how to brief and test a machine, and what must remain human. The knowledge is taught, applied in purposeful work and revisited with more responsibility.
One taught journey10 → 18
4age bandsfrom 10 to 18
10curriculum strandsin every band
7human capabilitiespracticed through work
Four age-shaped products connect knowledge, human capability and useful work. The curriculum grows with students; it is not a long staircase of disconnected tips.
4 age bands × 10 strands
The AI knowledge spine, made more demanding with age.
S1AI mental models
Prediction, models, limits and uncertainty.
S2Data and representation
Training data, classification, bias, omission and provenance.
S3Briefing and delegation
Define a job, inputs, outputs, limits and human approval.
S4Truth and evidence
Sources, corroboration, claims, hallucinations and recency.
S5Cognitive custody
Attempt first, smallest help, judgement and AI-off transfer.
S6Creation and craft
Make quality work across media and use specialist tools.
S7People and inclusion
Listen, design for access, examine power, benefit and harm.
S8Testing and revision
Run user, source, experiment, bias, access and edge-case tests.
S9Collaboration and leadership
Record decisions, roles, conflict, handover and accountability.
S10Venture and governance
Create value while governing operations, resources, AI and closure.
Safety, care and ethical judgment run through every age band. They are part of the interesting work, not a policy slide after it.
The cognitive-custody loop
The student remains responsible through all four moves.
The machine may contribute inside a declared boundary. It does not get the first word, the final decision or the explanation. That is how students use AI without quietly letting it do the thinking for them.
01STUDENT
Make firstRecord a real starting point and set the test.
02STUDENT
DelegateGive AI one bounded job—or choose AI Off.
03STUDENT
VerifyTest the exact contribution against evidence or reality.
04STUDENT
DefendMake the human call and explain it with AI closed.
The record on the workAI contributed · I decided · I can explain
The progression
From checking one answer to leading a whole system.
4–501
Spota visible error and ask “says who?”
6–702
Tracea claim, a source and the work’s provenance
8–1003
Testa serious build against evidence and real use
11–1204
Governpermissions, failure, redress and reruns
One verb spinespot → compare → trace → test → edit → constrain → ship → govern
See the journey by age
What changes at each age?
Choose a band to see the AI knowledge and skills students practice, what they make and what they can show.
Band 1 · ages 4–5
They meet AI as a pattern machine to supervise—not a person, an oracle or a shortcut.
Make & TestIllustrative learning routework they can make, test and explain
01AI knowledge + skill
- Tell an AI output from a source and ask “says who?”
- Keep private details and other people’s images out of prompts
- Choose a hint that leaves the thinking job with them
02They make
- A Make Me Get It Wrong challenge
- A source-backed reveal panel
- A checked Challenge Capsule and family AI agreement
03They can show
- The first call they made before the machine appeared
- The source that decided what survived
- What was mine, AI-helped and mine again
Example curriculum route
AThe machine and my jobBKeep people & pictures safeCCheck and askDHelp without takeoverEMake, hand over, explain
Governed curriculum release
The architecture is fixed. Every released package still has to earn its place.
Own Mind separates the curriculum promise from the content-release state. A package is provisioned only when its teaching, source, access, custody, safeguarding and school-control requirements are inspectable.
The progression, output families and deep-lesson standard describe the product schools are buying.
Released content must culminate in meaningful work, not a thin sequence of generated questions.
Managed identity, class scope, task contracts and evidence limits remain explicit before students begin.
This tour explains the curriculum architecture and shows illustrative route names. It does not publish an exact supplied-package inventory. Commercial scope is derived from the governed manifest, and schools approve every project version provisioned for their classes.
Starting later
Start at any age band—not back at the beginning.
A school can join at the band that fits its students. A new student enters through one age-respectful project, never a younger-band replay. No readiness score, placement decision, catch-up percentage or inferred profile is produced.
Age-right entry · Maker Desk
One real first make—never a babyish catch-up.
A student new at 4–5 starts with a bounded, real make for their own class: listen to a genuine need, make a first version, give AI one small job or none, test it with a real reader and explain the decision. There is no prerequisite from an earlier year.
Evidence carried forwardFirst idea · what AI did · reader test · change made · AI-off explanation
How a new student enters
- Listen to a real need and frame it
- Make a first version, then bound any machine job
- Test with a real reader, revise and explain with AI off
If the work exposes a precise gap, the teacher can add a short just-in-time bridge clinic. A returning Own Mind student can instead carry selected earlier work forward as context—useful evidence, never a gate.
No deficit label. No younger-band prerequisite. No automatic placement. Every band has an equal, age-respectful way in.
The 4–12 journey
From spotting a machine’s limits to governing a responsible workflow.
The same ideas return with greater independence, more demanding evidence and more consequential decisions. At every age, students know, do, make and defend.
Maker Desk
Make for someone.
Know
How people, rules, search and generative systems differ; why a claim needs evidence; what information must stay private; and the difference between known and unknown.
Do
Follow and improve a simple plan, compare a machine output with a source or user need, notice an access problem and explain one important choice.
Make
A usable guide, a checked image-and-claim, an access improvement, or a student-authored story or internal contribution.
Defend
Show the starting idea, one real test and one revision, then explain the work in plain language with AI closed.
Contributor Studio
Contribute with others.
Know
How provenance, source quality, classification, representation and omitted perspectives affect a conclusion; how roles and reversible tests protect a team decision.
Do
Compare sources, identify a missing voice, choose a bounded machine job, negotiate an explicit team decision and change work in response to evidence.
Make
A source investigation, accessible school guide, evidence-backed recommendation, team service or community contribution.
Defend
Identify the individual contribution, sources used, machine contribution, team decision and the feedback that changed the work.
Venture Studio
Improve a real system.
Know
How substantive domain knowledge, contested evidence, privacy, intellectual property, qualification rules and system trade-offs constrain responsible work.
Do
Frame a need, brief a bounded role, establish human approval points, test edge cases and bias risks, revise consequentially and stop when the evidence requires it.
Make
A tested product, service, recommendation, system improvement or small value-creating initiative for a real or credible user.
Defend
Separate student and machine contributions; present the exact test, consequential rework, return or handover; and answer questions with AI closed.
Venture HQ
Govern and operate something useful.
Know
How stakeholders, resources, data, evidence, risk and AI roles interact; when approval, monitoring, shutdown, disclosure, closure and handover are required.
Do
Run evidence-led experiments, set responsibilities and stop conditions, govern an AI workflow, make accountable commitments and close or hand over responsibly.
Make
A venture case, operating artifact, experiment record, AI-workforce governance file, board pack and handover or closure record.
Defend
Take individual responsibility before a board-style review, justify the next commitment from evidence and account for failure, shutdown and succession conditions.
AI literacy from 10 to 18
The same knowledge returns at greater depth.
Ten strands return in every band. Students move from recognizing and explaining, to choosing and applying, to testing and governing. Independence, complexity, evidence and consequence increase each time.
| Recurring strand | 4–5 · Maker Desk | 6–7 · Contributor Studio | 8–10 · Venture Studio | 11–12 · Venture HQ |
|---|---|---|---|---|
| 01AI mental modelsPrediction, models, limits and uncertainty. | Identify different kinds of system; predict a simple output; name a visible limit. | Compare models and outputs; explain a likely omission or failure. | Choose an appropriate role for a system and anticipate where it may fail. | Govern a system with monitoring, escalation and stop conditions. |
| 02Data and representationTraining data, classification, bias, omission and provenance. | Recognize training examples, categories and personal information. | Trace classification, representation, bias and provenance. | Justify data boundaries and test whose experience is missing. | Govern collection, access, retention, disclosure and deletion. |
| 03Briefing and delegationDefine a job, inputs, outputs, limits and human approval. | Give one bounded job with a clear output. | Specify inputs, limits, output and a human approval point. | Write and test a role contract with checks and failure handling. | Govern an AI-workforce contract, including shutdown and handover. |
| 04Truth and evidenceSources, corroboration, claims, hallucinations and recency. | Distinguish a claim from the evidence offered for it. | Check authority, currency, scope, corroboration and missing evidence. | Build a warranted case, address counterclaims and state uncertainty. | Set evidence thresholds, review routes and incident responses. |
| 05Cognitive custodyAttempt first, smallest help, judgement and AI-off transfer. | Make the first move and say what help, if any, the machine gave. | Hold an individually owned contribution inside shared work. | Protect the consequential decisions and complete an AI-off defense. | Remain the accountable human for commitments, exceptions and closure. |
| 06Creation and craftMake quality work across media and use specialist tools. | Guide, draft, prototype and improve something useful for a person. | Create an accessible service or team output to a stated standard. | Produce external-quality work or improve a real system. | Operate, maintain, hand over or responsibly close a useful artifact. |
| 07People and inclusionListen, design for access, examine power, benefit and harm. | Notice a user need and make one access improvement. | Seek missing perspectives and handle consent and representation. | Examine power, benefit, burden and exclusion in a proposed change. | Create accountable stakeholder, access and redress arrangements. |
| 08Testing and revisionRun user, source, experiment, bias, access and edge-case tests. | Run one exact check and change the work when it fails. | Use source, user and access tests; preserve what changed and why. | Test edge cases, bias, regression and version-specific behavior. | Run a test suite, monitor operation and trigger a safe stop. |
| 09Collaboration and leadershipRecord decisions, roles, conflict, handover and accountability. | Explain a choice, keep a role and return work to the right person. | Negotiate roles, record dissent and complete a usable handover. | Lead a decision, repair coordination and account for team choices. | Take board responsibility and plan succession or closure. |
| 10Venture and governanceCreate value while governing operations, resources, AI and closure. | Make something useful and return it responsibly. | Contribute to a small service and plan for continuity. | Test value, improve a system and close without abandoning obligations. | Govern resources, operations, risk, handover and responsible closure. |
AI mental models
Prediction, models, limits and uncertainty.
- 4–5 · Maker Desk
- Identify different kinds of system; predict a simple output; name a visible limit.
- 6–7 · Contributor Studio
- Compare models and outputs; explain a likely omission or failure.
- 8–10 · Venture Studio
- Choose an appropriate role for a system and anticipate where it may fail.
- 11–12 · Venture HQ
- Govern a system with monitoring, escalation and stop conditions.
Data and representation
Training data, classification, bias, omission and provenance.
- 4–5 · Maker Desk
- Recognize training examples, categories and personal information.
- 6–7 · Contributor Studio
- Trace classification, representation, bias and provenance.
- 8–10 · Venture Studio
- Justify data boundaries and test whose experience is missing.
- 11–12 · Venture HQ
- Govern collection, access, retention, disclosure and deletion.
Briefing and delegation
Define a job, inputs, outputs, limits and human approval.
- 4–5 · Maker Desk
- Give one bounded job with a clear output.
- 6–7 · Contributor Studio
- Specify inputs, limits, output and a human approval point.
- 8–10 · Venture Studio
- Write and test a role contract with checks and failure handling.
- 11–12 · Venture HQ
- Govern an AI-workforce contract, including shutdown and handover.
Truth and evidence
Sources, corroboration, claims, hallucinations and recency.
- 4–5 · Maker Desk
- Distinguish a claim from the evidence offered for it.
- 6–7 · Contributor Studio
- Check authority, currency, scope, corroboration and missing evidence.
- 8–10 · Venture Studio
- Build a warranted case, address counterclaims and state uncertainty.
- 11–12 · Venture HQ
- Set evidence thresholds, review routes and incident responses.
Cognitive custody
Attempt first, smallest help, judgement and AI-off transfer.
- 4–5 · Maker Desk
- Make the first move and say what help, if any, the machine gave.
- 6–7 · Contributor Studio
- Hold an individually owned contribution inside shared work.
- 8–10 · Venture Studio
- Protect the consequential decisions and complete an AI-off defense.
- 11–12 · Venture HQ
- Remain the accountable human for commitments, exceptions and closure.
Creation and craft
Make quality work across media and use specialist tools.
- 4–5 · Maker Desk
- Guide, draft, prototype and improve something useful for a person.
- 6–7 · Contributor Studio
- Create an accessible service or team output to a stated standard.
- 8–10 · Venture Studio
- Produce external-quality work or improve a real system.
- 11–12 · Venture HQ
- Operate, maintain, hand over or responsibly close a useful artifact.
People and inclusion
Listen, design for access, examine power, benefit and harm.
- 4–5 · Maker Desk
- Notice a user need and make one access improvement.
- 6–7 · Contributor Studio
- Seek missing perspectives and handle consent and representation.
- 8–10 · Venture Studio
- Examine power, benefit, burden and exclusion in a proposed change.
- 11–12 · Venture HQ
- Create accountable stakeholder, access and redress arrangements.
Testing and revision
Run user, source, experiment, bias, access and edge-case tests.
- 4–5 · Maker Desk
- Run one exact check and change the work when it fails.
- 6–7 · Contributor Studio
- Use source, user and access tests; preserve what changed and why.
- 8–10 · Venture Studio
- Test edge cases, bias, regression and version-specific behavior.
- 11–12 · Venture HQ
- Run a test suite, monitor operation and trigger a safe stop.
Collaboration and leadership
Record decisions, roles, conflict, handover and accountability.
- 4–5 · Maker Desk
- Explain a choice, keep a role and return work to the right person.
- 6–7 · Contributor Studio
- Negotiate roles, record dissent and complete a usable handover.
- 8–10 · Venture Studio
- Lead a decision, repair coordination and account for team choices.
- 11–12 · Venture HQ
- Take board responsibility and plan succession or closure.
Venture and governance
Create value while governing operations, resources, AI and closure.
- 4–5 · Maker Desk
- Make something useful and return it responsibly.
- 6–7 · Contributor Studio
- Contribute to a small service and plan for continuity.
- 8–10 · Venture Studio
- Test value, improve a system and close without abandoning obligations.
- 11–12 · Venture HQ
- Govern resources, operations, risk, handover and responsible closure.
Habits of mind
Seven human capabilities, taught one move at a time.
Each project concentrates on two or three capabilities. Students see the move modeled, rehearse it, use it independently and meet it again when the subject, team or problem changes. Across the year, all seven return.
Self-regulation and executive function
- Students learn
- Set a goal, sequence dependencies, begin, monitor, seek help, recover and hand over or close responsibly.
- How it is taught
- The course models a workable plan, exposes a dependency error and guides goal, checkpoint and recovery decisions before fading the prompts.
- You see it in
- Goal and plan · checkpoint decision · help request · recovery action · handover or closure
- It returns when
- Later work changes the constraints or interrupts the plan, requiring the student to recover without the original scaffold.
Metacognition
- Students learn
- Predict the demand, select a strategy, monitor understanding, diagnose an error, change approach and explain transfer.
- How it is taught
- Students compare strategies, see a modeled error diagnosis, state an intended approach and later explain why they retained or changed it.
- You see it in
- Strategy prediction · monitoring note · diagnosed error · changed approach · transfer explanation
- It returns when
- A changed task asks the student to select the strategy rather than repeat the one named in the original lesson.
Argument formation and critical reasoning
- Students learn
- Build and bound a position using claim, evidence, warrant, source quality, counter-position, uncertainty and limits.
- How it is taught
- The course models claim–evidence–warrant reasoning, contrasts strong and weak source use and requires a counter-position or evidential limit.
- You see it in
- Source comparison · warranted claim · correction · counter-position · defended limitation
- It returns when
- An unfamiliar source set changes what is credible, so the student must rebuild rather than reproduce the first argument.
Creative problem-framing
- Students learn
- Distinguish symptom from need, separate evidence from assumption, identify affected people, compare frames and revise the problem.
- How it is taught
- The course contrasts a superficial complaint with an evidenced need, guides a second case and requires an independently framed problem with observable criteria.
- You see it in
- Observation or user report · evidence/assumption split · initial frame · revised frame · criteria
- It returns when
- At least three weeks later, the student frames a new need without the original prompts or AI.
Collaboration and negotiation
- Students learn
- Listen and revoice, accept responsibility, surface disagreement, negotiate, decide, preserve dissent, repair and hand over.
- How it is taught
- Students use bounded roles, restate another view before responding, record a decision and show how feedback or disagreement changed the work.
- You see it in
- Role agreement · revoiced position · dissent and decision · feedback-led change · individual contribution · handover
- It returns when
- A later team uses different roles or a genuine disagreement, requiring the move to survive outside the original group routine.
Ethical reasoning and moral judgment
- Students learn
- Identify affected parties, distinguish facts from values, examine rights, duties, care, justice, benefit and burden, then justify acting, holding or stopping.
- How it is taught
- The course models a stakeholder trade-off, reveals a missing or burdened party and asks for a revisable decision with an explicit limit.
- You see it in
- Stakeholder map · fact/value distinction · trade-off · permission decision · act/hold/stop judgment · disclosure
- It returns when
- A changed stakeholder or consequence tests whether the student can reconsider the decision rather than defend it automatically.
Comfort with ambiguity and adaptive thinking
- Students learn
- Separate known, unknown and assumed; hold alternatives; run a reversible test; update after surprise; adapt, hold, pivot or stop.
- How it is taught
- Students predict what evidence would change their view, run a bounded test and revisit the decision when the result is incomplete or unexpected.
- You see it in
- Uncertainty record · alternatives · prediction · reversible test · changed decision · response to an unfamiliar case
- It returns when
- A later case with incomplete evidence asks the student to choose a responsible next move without premature certainty.
With course modeling
The course makes the move and its reasoning visible through an authored example and non-example.
With course guidance
The student rehearses it with prompts, feedback and a shared standard; the course responds to the current answer.
Independently
The student performs it in a familiar context without the instructional scaffold.
Adaptively
The student selects and adjusts it in a new or changed context after the original cues have gone.
Development across contexts
Progress becomes visible as support fades and situations change.
Each project preserves the first attempt, decision, test, revision and student explanation. The same capability returns in later projects and subject work. At checkpoints, a named professional compares selected work with shared progression anchors and records how independently the student now uses the move.
Focus
Each project selects two or three capabilities, names the exact move and creates a real decision in which students need to use it.
Teach
The course names the move, models it, guides rehearsal, checks understanding and fades support before independent use.
Capture
The first attempt, decision, work version, test, revision and student explanation stay connected to the context in which they happened.
Revisit
The same move returns after a delay and through the Learning Bridge, where a different subject or situation asks the student to adapt it.
Review
At a planned checkpoint, a named school professional connects selected work across contexts in a concise development statement.
Share
The statement is moderated against shared anchors and examples, then shown to the student alongside the work that supports it.
The measure is what students can do with less help, in a different context.
The shared progression anchor is with course modeling → with course guidance → independently in a familiar context → adaptively in a new or changed context.
A named school professional reviews selected work against shared examples, records a concise development statement and shows it to the student beside the supporting work.
Not worksheets. Work.
Seven families of things worth making.
Media is not the curriculum. A slide deck, a prototype and a service can all be weak work. The output family names the kind of value the student is trying to create and the responsibility it carries.
Useful thing
A physical, digital or mixed object someone can genuinely use.
Explanation or cultural work
A guide, model, exhibition, performance, story, archive or teaching resource.
Evidence-backed decision
A recommendation, options brief or bounded conclusion that follows from evidence.
Service
A repeatable act of value with an owner, operating notes and a responsible close.
Improved system
A tested change to a school or community process.
Community contribution
Work shaped by listening, permission, feedback, revision and return.
Own learning
A student-directed inquiry, learning asset, project or venture inside a teacher-governed envelope.
See the complete Maker and project story, including the creation routes and named project examples →
A record made from real work
See the thinking develop, not just the finished answer.
First attempt. Chosen strategy. Sources. Bounded machine contribution. Test. Revision. Independent explanation. Together, these show how understanding becomes work—and whether it travels later.
A check opens the right next step.
Every student answers an unmarked hinge question and explains the reasoning before seeing feedback. The course then releases the authored reteach, scaffold or extension branch.
Thinking stays connected to the work.
The record preserves sources, a substantive first version, the machine role, individual and team decisions, the exact test, consequential revision and return or closure.
Knowledge returns without the original prompts.
A shorter independent task with AI closed asks students to retrieve important knowledge and use it in a changed case.
The Learning Bridge changes the context.
A subject task or unfamiliar project asks the student to select and adapt the relevant knowledge and habit for themselves.
Learning design
Built for understanding that lasts and travels.
Worked examples lead into guided practice and independent performance. Retrieval brings important knowledge back after a delay. Varied tasks reveal transfer. Criteria and exemplars help students recognize quality and improve their own work.
Worked examples, self-explanation and fading
A complete example leads into prompted explanation, a partially completed second case and independent performance.
Atkinson, Renkl & Merrill (2003)Metacognition in meaningful curriculum work
Planning, monitoring and evaluation are taught inside substantive tasks, with prompts fading as students take control.
Education Endowment Foundation (2025)Spacing and retrieval practice
Important knowledge returns through low-stakes retrieval distributed across later lessons and projects.
Carpenter, Pan & Butler (2022)Transfer under changed conditions
Later work changes the brief, subject, tool, evidence or constraints so students must select and adapt what they learned.
Pan & Rickard (2018)Criteria, exemplars and self-monitoring
Students compare work with clear criteria and exemplars, then use specific moves to revise; reviewers moderate against shared anchors.
Sadler (1989)AI literacy for responsible participation
The AI-literacy spine connects human-centered judgment, ethics, techniques and applications, and system design.
UNESCO AI Competency Framework for Students (2024)Safety and cognitive development in education AI
Age, data, transparency, progressive disclosure, filtering, monitoring and human escalation shape every approved AI role.
Department for Education product safety standards (2026)Inside every lesson
Explanation, practice, feedback and transfer are built into the course.
The course explains and models new ideas, guides practice, checks understanding and releases the next authored pathway. Students make, test, revise and defend the work. Schools timetable the program, supervise safe collaboration, support access and review selected work.
Support fades as responsibility grows.
Worked examples, guided rehearsal, hinge questions and authored reteach, scaffold or extension routes.
The work shows where the thinking happened.
A first attempt, consequential decisions, exact test, revision, disclosure and independent defense.
Time, safety and professional review.
Protected curriculum time, supervised collaboration, access support, safeguarding and moderated review of selected work.
The design principle
Every released lesson must leave the student knowing more—and holding work they can stand behind.
Curriculum + Make
Give AI and human capability a serious place on the timetable.
Start with a complete dedicated program. Add Across Subjects later if and when your school wants the same practice to travel into disciplinary work.