About

Every child should be able to understand AI, not only use it.

Why we’re doing this

Children already meet AI in the apps, games and search tools they use. Most of the time nobody explains how it works. We think understanding should come first: how AI learns from examples, why it makes mistakes, and who is responsible when it gets things wrong.

Wondi teaches these ideas through short, hands-on lessons for ages 5 to 14. We want children to be curious, not scared, and confident enough to question what an AI tells them.

Our aim

To bring families the most cost-effective and efficient way to gain AI knowledge.

Our principles

Curious, not scared. Benefits and risks, taught evenly.
Humans in charge. People design, test, decide and take responsibility.
Data and fairness at the centre. The examples a machine learns from matter.
Privacy first. We collect as little as we can, only from parents, and we follow the spirit of the ICO’s Children’s Code.

Who runs this site

Adcore Edge Solutions FZCO.

Get in touch: wondi@adcoreedge.com.

How Wondi fits school guidelines

We mapped each course to what England, the UAE, India and the United States ask children to learn about AI and computing. Last checked October 2026.

This is our own reading of public documents. It is not an approval, accreditation or endorsement by any government, ministry or exam board. Ages and grades are approximate, and schools differ. Check the sources with your school.

UK England (UK)National Curriculum: Computing

A1 Explorersages 5 to 7Key Stage 1 Partial fit
A2 Buildersages 7 to 9Key Stage 2 Partial fit
A3 Creatorsages 9 to 11Key Stage 2 Partial fit
A4 Codersages 11 to 14Key Stage 3 Strong fit

UAE United Arab EmiratesUAE National AI Literacy Curriculum Framework (K to 12)

A1 Explorersages 5 to 7Cycle 1 Strong fit
A2 Buildersages 7 to 9Cycle 1 Strong fit
A3 Creatorsages 9 to 11Cycle 1 to 2 Strong fit
A4 Codersages 11 to 14Cycle 2 to 3 Strong fit

India IndiaNEP 2020 and NCF-SE 2023

A1 Explorersages 5 to 7Before Class 3 Light fit
A2 Buildersages 7 to 9Classes 3 to 4 Partial fit
A3 Creatorsages 9 to 11Classes 5 to 6 Partial fit
A4 Codersages 11 to 14Classes 6 to 9 Strong fit

US United States2026 CSTA PK–12 Computer Science Standards and the AI4K12 Five Big Ideas (voluntary; each state sets its own standards)

A1 Explorersages 5 to 7Grades K to 2 Strong fit
A2 Buildersages 7 to 9Grades 2 to 4 Strong fit
A3 Creatorsages 9 to 11Grades 4 to 6 Partial fit
A4 Codersages 11 to 14Grades 6 to 9 Partial fit

Strong fit Partial fit Light fit (builds readiness)

UK England (UK)National Curriculum: Computing, with a new curriculum being drafted

What it asks of children

The current Computing programme of study has no AI content. The government has confirmed that Computing stays the home for digital literacy, critical thinking and foundational AI concepts. A new curriculum is being drafted, with a public consultation expected later in 2026, and the government’s timetable points to first teaching from September 2028.

Key Stage 1 (ages 5 to 7): know what algorithms are, use technology safely and keep personal information private.
Key Stage 2 (ages 7 to 11): explain simple algorithms and find errors, use search and judge digital content critically, and use technology responsibly.
Key Stage 3 (ages 11 to 14): design, use and evaluate computational models, use a textual programming language, and protect online identity and privacy.

How Wondi lines up

Wondi course	Stage	Fit	What lines up
A1 Smart or Not?Explorers, ages 5 to 7	Key Stage 1 (Years 1 to 2)	Partial fit	Algorithms: machines follow precise instructions, people do other things better (A1 L1). Online safety: say what information is private and make safe-use rules (A1 L5, L6). Digital literacy: spot technology in everyday life (A1 L2). Digital literacy: design a smart helper and say who checks it (A1 L7). Readiness, beyond Key Stage 1.
A2 Teach a MachineBuilders, ages 7 to 9	Lower Key Stage 2 (Years 3 to 4)	Partial fit	Logical reasoning: count right and wrong results, find errors and improve a model (A2 L4). Data: collect examples and present results (A2 L2, L6). Digital literacy: judge whether examples are fair and complete (A2 L5). Digital literacy: where AI is used beyond school and who stays in charge (A2 L7).
A3 How Machines LearnCreators, ages 9 to 11	Upper Key Stage 2 (Years 5 to 6)	Partial fit	Evaluating digital content: simple checks for reliable, fake or computer-made content (A3 L5). Logical reasoning: calculate accuracy and describe the kinds of mistakes (A3 L4). Data: build and test a classifier with your own data (A3 L2, L3). Responsible use: say where AI can fail and how to use it wisely (A3 L6). Responsible use: plan an AI helper, say who checks it and how it could be unfair (A3 L7).
A4 AI in PracticeCoders, ages 11 to 14	Key Stage 3 (Years 7 to 9)	Strong fit	Computational models: train, test and evaluate a model in Python (A4 L1 to L3). Data: features, labels, bias in data and its effects (A4 L1, L4). Trustworthiness: why language models give confident wrong answers (A4 L5). Online safety: privacy, copyright and a personal code of conduct (A4 L6). Responsible use: AI across sectors, who is accountable, and a reasoned argument (A4 L7). The Local guidance box points to DfE guidance for families with an England account.

Assessment and school safety, lesson by lesson

England: assessment and school safety, lesson by lesson

In England there is no national test for Computing at Key Stages 1 to 3. Teachers assess against the programme of study, so each lesson shows the evidence a teacher can use. DfE guidance says AI must not be the sole basis for an assessment decision, and Wondi does not make one: the child's work and the teacher's judgement do.

A1 Smart or Not? (ages 5 to 7, Key Stage 1 (Years 1 to 2))

Smart or Not?: lesson evidence against the England Computing programme
Lesson	Computing programme of study	Evidence the child produces	Assessment	Fit
L1 Machines and People	Algorithms: machines follow precise instructions	Card sort: machine good, person good, both.	Formative	Meets
L2 Helpers Around Us	Digital literacy: spot technology in everyday life	Ring the smart helpers on a picture page.	Formative	Meets
L3 Sorting Game	Algorithms: following and learning from examples	Train a partner-robot with 2 then 6 examples.	Formative	PartlyLearning from examples is beyond the Key Stage 1 programme, so this is readiness.
L4 Robots Get It Wrong	Logical reasoning: check answers, find mistakes	Fix the wrong answers in picture pairs.	Formative	Meets
L5 Be Kind to Data	Online safety: say what is private	Sort cards into fine to share and keep private.	Formative	Meets
L6 Our Rules for Smart Helpers	Online safety: safe-use rules	Poster with 3 safety rules and a picture quiz.	Summative	Meets
L7 Design a Smart Helper	Digital literacy: use technology purposefully to create and organise ideas	Drawing of a smart helper labelled with its job, what it looks at and who checks it.	Formative	PartlyDesigning an AI helper goes beyond the Key Stage 1 programme, so this is readiness.

A2 Teach a Machine (ages 7 to 9, Lower Key Stage 2 (Years 3 to 4))

Teach a Machine: lesson evidence against the England Computing programme
Lesson	Computing programme of study	Evidence the child produces	Assessment	Fit
L1 What Is AI?	Algorithms: what a program is, and how AI differs	One sentence: AI is…	Formative	PartlyAI itself is not in the current programme, so this goes beyond it.
L2 Examples Teach Machines	Data: collect and use examples	Say why the examples matter.	Formative	Meets
L3 Train a Picture Sorter	Data: sort and group examples	A classifier trained with at least 2 classes.	Formative	PartlyThe classifier tool is still in development. Use drawn example cards until it ships.
L4 Test It	Logical reasoning: find errors and improve	Count right and wrong results and improve the model.	Formative	Meets
L5 When Examples Are Unfair	Digital literacy: judge whether content is fair and complete	Spot examples that leave things out.	Formative	Meets
L6 Show and Share	Data: present results	Short presentation of the model and where it may fail.	Summative	Meets
L7 AI at Work Around Us	Digital literacy: uses of technology beyond school	Job cards sorted into AI can help, people decide and both.	Formative	Meets

A3 How Machines Learn (ages 9 to 11, Upper Key Stage 2 (Years 5 to 6))

How Machines Learn: lesson evidence against the England Computing programme
Lesson	Computing programme of study	Evidence the child produces	Assessment	Fit
L1 Rules Versus Learning	Algorithms: explain simple algorithms	Explain rules versus learning with an example.	Formative	Meets
L2 Training and Testing Data	Data: test with fresh examples	Say why a model needs unseen test examples.	Formative	Meets
L3 Build a Classifier	Computational thinking: design, build and test	A classifier built and tested on the child's own data.	Formative	PartlyThe classifier tool is still in development.
L4 Accuracy and Mistakes	Logical reasoning: find and describe errors	Accuracy worked out and the kinds of mistakes described.	Formative	Meets
L5 Real, Fake or Made-up?	Digital literacy: judge digital content critically	Apply simple checks to real, fake and made-up examples.	Formative	Meets
L6 Using AI Wisely	Online safety: use technology responsibly; identify ways to report concerns	One-page how it works and where it fails, with responsible use.	Summative	PartlyResponsible use is taught. Reporting concerns is not, so add a short teacher-led step on who to tell.
L7 AI for Good, AI with Rules	Online safety and digital literacy: use technology responsibly and judge it	Group planning sheet for an AI helper, with who checks it and one way it could be unfair.	Formative	PartlyAI for good is not in the current programme, so this goes beyond it.

A4 AI in Practice (ages 11 to 14, Key Stage 3 (Years 7 to 9))

AI in Practice: lesson evidence against the England Computing programme
Lesson	Computing programme of study	Evidence the child produces	Assessment	Fit
L1 How Models Predict	Computational models of real-world problems	Explain how features lead to a prediction.	Formative	PartlyThe programme means models and simulations of real-world systems. A trained classifier is a close reading, not the same thing, and children explain a model here rather than design one.
L2 Build and Train a Classifier	Textual programming language	A simple classifier trained in Python.	Formative	PartlyPython is used only to train a classifier. It does not cover general programming.
L3 Training, Testing and Over-fitting	Evaluate computational models	A fair test split and an over-fitting explanation.	Formative	Meets
L4 Bias in Data and Decisions	Evaluate models: bias in data	Bias found in a dataset and its effect described.	Formative	Meets
L5 Language Models and Made-up Answers	Digital literacy: trust and reliability of content	Explain why language models give confident wrong answers.	Formative	Meets
L6 Ethics, Privacy, Copyright and My AI Code	Online safety: protect identity and privacy; recognise inappropriate content, contact and conduct; report concerns	A reasoned argument and a personal code of conduct.	Summative	PartlyPrivacy and responsible use are taught. Recognising inappropriate contact and conduct, and how to report concerns, are not.
L7 AI, Society and Rules	Digital literacy: use technology safely, respectfully and responsibly	Sector table and a two-minute argument on whether a use of AI should be encouraged, limited or stopped.	Summative	PartlyThe Local guidance box points to DfE guidance, but it summarises it briefly and has not been checked against the current document.

The Computing programme of study, statement by statement

The GOV.UK National curriculum collection has 18 documents and none is about AI. The Computing programme of study is dated 11 September 2013. The other subject documents (maths, science, design and technology, citizenship) and the RSHE statutory guidance are not mapped here, because we could not read their full text. They are the next place to check, since AI literacy can sit in data handling, evaluating evidence and online safety.

England Computing programme of study statements and where Wondi covers them
Statement	Fit	Where
Key Stage 1: Algorithms are precise instructions that run as programs	Meets	A1 L1
Key Stage 1: Create and debug simple programs, and predict what they do	Gap	No lesson. Wondi does not teach programming.
Key Stage 1: Use technology to create, store and retrieve content, and recognise everyday uses	Partly	A1 L2 covers everyday uses only.
Key Stage 1: Online safety: keep personal information private and know where to get help	Meets	A1 L5, L6
Key Stage 2: Design, write and debug programs using sequence, selection, repetition and variables	Gap	No lesson.
Key Stage 2: Explain simple algorithms and detect and correct errors	Meets	A2 L4, A3 L1, L4
Key Stage 2: Understand computer networks and use search technologies effectively	Partly	A3 L5 covers judging content. Networks are not taught.
Key Stage 2: Collect, analyse, evaluate and present data and information	Meets	A2 L2, L6, A3 L2 to L4
Key Stage 2: Online safety: use technology safely, respectfully and responsibly, and identify ways to report concerns	Partly	A3 L6 covers responsible use. Reporting concerns is not taught.
Key Stage 3: Design, use and evaluate computational models of real-world problems	Partly	A4 L1 to L3 use classifiers, which are close to but not the same as models and simulations of real-world systems.
Key Stage 3: Understand key algorithms such as sorting and searching	Gap	No lesson.
Key Stage 3: Use two programming languages, one textual, and write modular programs	Partly	A4 L2 uses Python only to train a classifier.
Key Stage 3: Understand Boolean logic and binary	Gap	No lesson.
Key Stage 3: Understand hardware, software, and how instructions and data are stored and processed	Gap	No lesson.
Key Stage 3: Collect and analyse data to meet the needs of known users	Partly	A4 L2 to L4 analyse data. Known users are not part of the project.
Key Stage 3: Online safety: protect online identity and privacy; recognise inappropriate content, contact and conduct; report concerns	Partly	A4 L6 covers privacy. Contact, conduct and reporting are not taught.

DfE product safety expectations for generative AI

Pupil-facing AI needs particular care. These 13 expectations were updated in January 2026. Our list comes from a summary of the DfE text, so check the DfE document itself before relying on it.

DfE generative AI product safety expectations against Wondi and Mavo
Expectation	Fit	How Wondi and Mavo line up
1. Stated purpose	Partly	Age groups, lessons and purpose are stated for every course. We have no impact evidence yet, so claims stay modest.
2. Educational use cases	Gap	We have not yet mapped Mavo to the DfE's named use cases.
3. Filtering	Meets	Built in, not bolted on. Every message is checked before the AI sees it, by pattern checks and an AI check, and every reply is checked again. Anything that is not a clear yes is blocked. No outside test yet.
4. Monitoring and reporting	Partly	A parent can read every chat for 30 days, and stopped messages are flagged. There is no alert to a school safeguarding lead, because Mavo is sold to families, not run by a school.
5. Security	Partly	Rules against jailbreaks, signed single-use sign-in links and a daily message cap. No independent security test and no check against the Cyber Security Standards for Schools and Colleges yet.
6. Privacy and data protection	Partly	Personal details are blocked from chats, parents can delete a chat or a family, and chats expire after 30 days. We have not published a school data protection pack or DPIA template, or said where data is processed.
7. Intellectual property	Partly	We do not use children's work to train models. The privacy notice should say so in plain words, and supplier terms should be confirmed.
8. Design and testing	Partly	Automated safety tests run on every build. We have not yet tested with children and teachers.
9. Governance	Gap	No published complaints route for schools or governance policy yet.
10. Cognitive development	Partly	Short answers, a daily message cap, and prompts to check answers in a second place. There is no hints-first design yet, so the AI may still give a full answer.
11. Emotional and social development	Meets	The AI says plainly that it is an AI. It must never pose as a human, a friend who needs the child or a family member, and it writes no romance or roleplay.
12. Mental health	Partly	Self-harm language is stopped and the child is pointed to a trusted grown-up, and a parent can see it. There is no school pathway, and we have no mental health expert on our governance.
13. Manipulation	Partly	A daily cap limits use. We have not yet reviewed rewards and badges against the rule on dark patterns, and the AI has no written rule against flattery.

What a school or trust will ask of us

Taken from the DfE guidance and a union model AI policy for schools.

What a school expects of an AI supplier
A school will ask for	Fit	Where we are
On the school's approved AI tools register, after a DPIA and supplier due diligence	Gap	We have no school pack yet: DPIA input, data flow, data location and retention in one place.
Due diligence on pedagogy, accessibility, bias, security, data location, deletion rights and value	Partly	Pedagogy, bias and deletion are covered by the lessons and by parent deletion. Accessibility and data location are not yet documented.
Contract: no training on school data, audit rights, deletion on termination	Gap	There is no school contract or data processing agreement yet.
Organisational accounts, not free consumer accounts	Partly	Family accounts and a separate classroom sign-in exist. A full school or trust account does not.
Personal data kept out of AI tools	Meets	Names, schools, addresses and other private details are blocked from chats and from the AI.
Not described as an AI or virtual teacher	Meets	Mavo is a supervised learning guide that says it is an AI. The lessons are teacher-led, and no page calls it a tutor or teacher.
A human-only pathway always available; staff never required to use AI	Meets	All four courses run unplugged, with no screen. Mavo is an optional add-on.
No reliance on AI detectors; AI is never the sole basis for an assessment decision	Meets	Wondi has no AI detector, and no AI marking of children's work.
Workload neutral, with a pilot and success criteria	Partly	Lesson plans are ready to teach, which should reduce preparation. We have not measured teacher workload in a pilot.

Not covered yet. England's current Computing programme has no AI content, and a new curriculum is planned for first teaching from September 2028. Wondi complements Computing and does not replace it. Wondi does not teach programming, key algorithms, Boolean logic, binary, networks or hardware, so it cannot stand in for the Computing programme. Mavo is a family product. Schools will need a school pack, a data processing agreement, an alert route to the safeguarding lead and a stated data location before adopting it. DfE says pupil-facing AI needs particular care and that most freely available tools are not suitable for pupils, so a school should run Mavo only with close supervision. Scotland, Wales and Northern Ireland have their own curricula, which are not mapped here.

How Wondi complements the gaps

How Wondi complements the England gaps
Gap	The school covers it with	Wondi today	Planned, subject to funding
Programming and debugging (Key Stages 1 and 2)	The school's Computing scheme, for example the National Centre for Computing Education's Teach Computing resources.	The unplugged lessons build the ideas that programs follow precise instructions and that errors are found and fixed (A1 L1, A3 L1, A3 L4).	A short lesson that links a program with fixed rules to a model that learns.
Key algorithms, Boolean logic, binary and hardware (Key Stage 3)	The school's Computing lessons.	A4 teaches how a model learns from data and is tested (L1 to L3), which sits beside the algorithms strand.	Teacher notes that show where each A4 lesson links to the Computing programme.
Reporting concerns online (Key Stages 2 and 3)	PSHE, relationships and health education, and the school's online safety lessons.	Children learn to ask a trusted grown-up first (A1 L5, L6) and write their own code of conduct (A4 L6).	A short teacher-led step on who to tell, added to A3 L6 and A4 L6.
Models and simulations of real-world systems (Key Stage 3)	The school's Computing lessons on modelling.	A4 L1 to L3 teach how a classifier is trained, tested and judged.	An extension where pupils test how a model behaves when its data changes.
Classifier tool for A2 to A4 (in development)	Not applicable.	Lesson plans are complete. Drawn example cards stand in, and the Try page shows the idea on screen.	The in-browser classifier, so pictures and sounds stay on the device.
No AI content in the current programme	Schools can teach AI ahead of the new curriculum in Computing, PSHE or assemblies.	Four ready-to-teach AI courses for ages 5 to 14, with evidence for each lesson.	A re-map to the new curriculum when it is published, with first teaching planned for September 2028.

Gaps and sources

What Wondi does not cover. Wondi does not teach general programming, networks or hardware. It complements the Computing curriculum and does not replace it. Scotland, Wales and Northern Ireland have their own curricula, which are not mapped here. Schools will also ask about DfE safety expectations for AI products, a data protection assessment and a school data agreement. Our lesson-by-lesson check and that list are in uk-assessment.mjs.

Sources

UAE United Arab EmiratesUAE National AI Literacy Curriculum Framework (K to 12), 2026

What it asks of children

The Ministry of Education made AI a subject in public schools from kindergarten to Grade 12 from the 2025 to 2026 school year. The framework has three domains, seven strands and three cycles: Cycle 1 (Grades K to 4), Cycle 2 (Grades 5 to 8) and Cycle 3 (Grades 9 to 12).

Cycle 1: explain that AI is a human-made tool that follows instructions and uses data, use AI safely and fairly, and spot everyday uses.
Cycle 2: explain how AI learns from data and uses algorithms, evaluate how well a model works, and explain how ethics and inclusion guide AI decisions.
Cycle 3: analyse how data affects AI decisions, evaluate AI governance, and design AI solutions to real-world problems.

How Wondi lines up

Wondi course	Stage	Fit	What lines up
A1 Smart or Not?Explorers, ages 5 to 7	Cycle 1 (KG to Grade 2)	Strong fit	AI is a human-made tool: machines and people (A1 L1) and smart helpers around us (A1 L2). Learning from examples: a sorting game (A1 L3). Safe and fair use: smart helpers make mistakes, private information, our three rules (A1 L4 to L6). Design a smart helper for a job at home and say who checks it (A1 L7).
A2 Teach a MachineBuilders, ages 7 to 9	Cycle 1 (Grades 2 to 4)	Strong fit	What AI is, in your own words (A2 L1). Data and machine learning: why examples matter, train and test a classifier (A2 L2 to L4). Fairness: spot examples that leave things out (A2 L5). Project: present a model and where it might fail (A2 L6). AI at work in different jobs, and who stays in charge (A2 L7).
A3 How Machines LearnCreators, ages 9 to 11	Cycle 1 to 2 (Grades 4 to 6)	Strong fit	Rules versus learning, training data versus test data (A3 L1, L2). Evaluate model performance: accuracy and mistakes (A3 L4). Judge real, fake and made-up content (A3 L5) and use AI wisely (A3 L6). Project: a mini AI project and a one-page “how it works and where it fails” (A3 L3, L6). AI for good, with rules: plan a helper, say who is responsible and whose values it respects (A3 L7).
A4 AI in PracticeCoders, ages 11 to 14	Cycle 2 to 3 (Grades 6 to 9)	Strong fit	How models predict, training, testing and over-fitting (A4 L1 to L3). Bias in data and decisions (A4 L4). Language models, prompts and made-up answers (A4 L5). Ethics, privacy and copyright, and a personal code of conduct (A4 L6). AI across sectors, accountability and a reasoned argument, with a Local guidance box that points to the UAE framework and Ministry guide (A4 L7). New and not yet classroom-tested.

UAE strands, course by course

The UAE framework has seven strands. This shows which Wondi courses cover each one, and in which lessons.

UAE AI literacy strands covered by each Wondi course
UAE strand	A1 Explorers	A2 Builders	A3 Creators	A4 Coders
1. Foundations and fundamentals of AI	StrongL1, L2	StrongL1	StrongL1	SomeL1
2. Data, algorithms and machine learning	SomeL3	StrongL2 to L4	StrongL2 to L4	StrongL1 to L3
3. Ethics, bias and fairness	SomeL4, L5, L7	StrongL5	SomeL4, L6	StrongL4, L6
4. AI policies and accountability	SomeL5, L6	SomeL7	SomeL6, L7	StrongL6, L7
5. AI applications in society and industry	StrongL2	SomeL7	SomeL5, L7	SomeL5, L7
6. Usage of AI tools	Somegrown-up-led AI tests	Somegrown-up-led AI tests	SomeL5 to L7	StrongL5
7. Innovation and project design	Someposter project, L7	StrongL6	StrongL3, L6	Strongproject and report

Assessment evidence, lesson by lesson

UAE assessment, lesson by lesson

The framework says assessment is school-based and gives principles and example tasks, but no rubrics. So each lesson shows the evidence a teacher can use. Principles: Understanding over memorising; Critical thinking over following steps; Ethics and safety alongside technical skill; Application and reflection.

A1 Smart or Not? (ages 5 to 7, Cycle 1 (KG to Grade 2))

Smart or Not?: lesson evidence against the UAE framework
Lesson	Strand	Framework standard	Evidence the child produces	Assessment	Fit
L1 Machines and People	1 Foundations and fundamentals of AI	Compare how AI is similar to or different from humans.	Picture-card sort: machine good, person good, both.	Formative: Reflection discussion	Meets
L2 Helpers Around Us	1 Foundations and fundamentals of AI; 5 AI applications in society and industry	Identify everyday uses of AI across different sectors.	Ring the smart helpers on a picture page.	Formative: Drawings and simple tasks	Meets
L3 Sorting Game	2 Data, algorithms and machine learning	Explain how AI systems use data to learn.	Train a partner-robot with 2 then 6 examples and compare.	Formative: Hands-on project	Meets
L4 Robots Get It Wrong	3 Ethics, bias and fairness	Evaluate when AI outputs can be trusted and when human judgement is required.	Fix the wrong answers in picture pairs and say who checks.	Formative: Reflection discussion	Meets
L5 Be Kind to Data	4 AI policies and accountability	Identify ways humans control AI to ensure fairness and protect personal information.	Sort cards into fine to share and keep private.	Formative: Simple task	Meets
L6 Our Rules for Smart Helpers	3 Ethics, bias and fairness; 7 Innovation and project design	Identify safe and fair ways to use AI. Design AI projects that support humans in everyday tasks.	Poster with 3 safety rules, gallery walk, picture quiz.	Summative: Drawing and short presentation	PartlyChildren make safety rules, not an AI helper design, so the design standard is only touched.
L7 Design a Smart Helper	7 Innovation and project design; 2 Data, algorithms and machine learning; 3 Ethics, bias and fairness	Design AI projects that support humans in everyday tasks. Explain how AI systems use data to learn. Identify safe and fair ways to use AI.	Drawing of a smart helper labelled with its job, what it looks at and who checks it, plus a partner talk on who it might leave out.	Formative: Drawings and simple tasks	Meets

A2 Teach a Machine (ages 7 to 9, Cycle 1 (Grades 2 to 4))

Teach a Machine: lesson evidence against the UAE framework
Lesson	Strand	Framework standard	Evidence the child produces	Assessment	Fit
L1 What Is AI?	1 Foundations and fundamentals of AI	Explain that AI systems are human-created tools that follow instructions and use data.	One sentence: AI is…, and a match of tools to rules or examples.	Formative: Simple task and question	Meets
L2 Examples Teach Machines	2 Data, algorithms and machine learning	Explain how AI systems use data to learn.	Say why the examples matter, with good and bad example sets.	Formative: Reflection discussion	Meets
L3 Train a Picture Sorter	2 Data, algorithms and machine learning; 7 Innovation and project design	Explain how AI systems use data to learn.	A classifier trained with at least 2 classes.	Formative: Hands-on project	PartlyThe classifier tool is still in development. Until it ships, use drawn example cards.
L4 Test It	2 Data, algorithms and machine learning	Evaluate AI model performance. (Cycle 2 standard, so this is ahead.)	Count right and wrong results and improve the model.	Formative: Hands-on project	Meets
L5 When Examples Are Unfair	3 Ethics, bias and fairness	Identify safe and fair ways to use AI. Grade 3: describe what bias means in AI.	Spot examples that leave things out and say who is missed.	Formative: Reflection discussion	Meets
L6 Show and Share	7 Innovation and project design; 3 Ethics, bias and fairness	Design AI projects that support humans in everyday tasks.	Short presentation: my model, how well it works, where it may fail.	Summative: Short presentation	Meets
L7 AI at Work Around Us	5 AI applications in society and industry; 4 AI policies and accountability	Identify everyday uses of AI across different sectors. Identify ways humans stay in control of AI.	Job cards sorted into AI can help, people decide and both, with who is in charge if it goes wrong.	Formative: Simple task and reflection discussion	Meets

A3 How Machines Learn (ages 9 to 11, Cycle 1 to 2 (Grades 4 to 6))

How Machines Learn: lesson evidence against the UAE framework
Lesson	Strand	Framework standard	Evidence the child produces	Assessment	Fit
L1 Rules Versus Learning	2 Data, algorithms and machine learning	Explain how AI systems learn from data and use algorithms to make decisions.	Explain rules versus learning with an example.	Formative: Quiz or spoken question	Meets
L2 Training and Testing Data	2 Data, algorithms and machine learning	Explain how AI systems learn from data and use algorithms to make decisions.	Say why a model needs unseen test examples.	Formative: Quiz or spoken question	Meets
L3 Build a Classifier	2 Data, algorithms and machine learning; 7 Innovation and project design	Design AI projects that address real-world problems or challenges.	A classifier built and tested on the child's own data.	Formative: Applied design project	PartlyThe classifier tool is still in development. The project is a small classroom problem, not a wider real-world one.
L4 Accuracy and Mistakes	2 Data, algorithms and machine learning	Evaluate AI model performance.	Accuracy worked out and the kinds of mistakes described.	Formative: Case or scenario analysis	Meets
L5 Real, Fake or Made-up?	3 Ethics, bias and fairness	Grade 5: identify the difference between real and AI-generated content. Explain how AI manipulates or creates media that may cause safety and security concerns.	Apply simple checks to real, fake and made-up examples.	Formative: Case or scenario analysis	Meets
L6 Using AI Wisely	3 Ethics, bias and fairness; 4 AI policies and accountability; 7 Innovation and project design	Identify how AI impacts privacy, accountability and the protection of personal rights.	One-page how it works and where it fails, with how I would use AI responsibly.	Summative: Portfolio with reflection and presentation	PartlyPrivacy and responsible use are covered. Accountability and national policy are not.
L7 AI for Good, AI with Rules	5 AI applications in society and industry; 4 AI policies and accountability; 7 Innovation and project design	Describe how AI is used for social and environmental good. Explain who is responsible for AI and whose values it should respect. Plan an AI project for a real need.	Group planning sheet (data, a way it could be unfair, who checks it, whose values) and a one-minute talk.	Formative: Group project plan and short presentation	PartlyChildren plan a helper but do not build or test one, and the global challenges in the framework are touched through two examples only.

A4 AI in Practice (ages 11 to 14, Cycle 2 to 3 (Grades 6 to 9))

AI in Practice: lesson evidence against the UAE framework
Lesson	Strand	Framework standard	Evidence the child produces	Assessment	Fit
L1 How Models Predict	1 Foundations and fundamentals of AI; 2 Data, algorithms and machine learning	Explain how advanced AI models process information.	Explain how features lead to a prediction.	Formative: Quiz or spoken question	PartlyFeatures and prediction are covered. Neural networks are not, and Grade 5 asks for the main parts of one.
L2 Build and Train a Classifier	2 Data, algorithms and machine learning	Explain how AI systems learn from data and use algorithms to make decisions.	A simple classifier trained on a dataset in Python.	Formative: Applied design project	Meets
L3 Training, Testing and Over-fitting	2 Data, algorithms and machine learning	Evaluate AI model performance.	A fair test split and an over-fitting explanation.	Formative: Case or scenario analysis	Meets
L4 Bias in Data and Decisions	3 Ethics, bias and fairness	Explain how ethics, cultural values and inclusion guide AI decision making.	Bias found in a dataset and its effect described.	Formative: Case or scenario analysis	Meets
L5 Language Models and Made-up Answers	2 Data, algorithms and machine learning; 3 Ethics, bias and fairness; 6 Usage of AI tools	Analyse how AI systems generate new outputs using patterns learned from data. Use generative AI models to create new content.	Explain next-word prediction and why answers can be confidently wrong.	Formative: Prompt check with feedback	PartlyUnderstanding is covered. Creating content with a generative tool is only done with a grown-up-led test or in Mavo, and the Ministry's Safe and Responsible Use Guide (Grades 7 to 12) governs which tools are approved.
L6 Ethics, Privacy, Copyright and My AI Code	3 Ethics, bias and fairness; 4 AI policies and accountability; 7 Innovation and project design	Identify how AI impacts privacy, accountability and the protection of personal rights. Examine how AI is governed through national policies.	A reasoned argument from several sides and a personal code of conduct.	Summative: Portfolio with reflection and ethical case study	PartlyPrivacy, copyright and ethics are covered. UAE national AI policy is not named, so this needs a teacher add-on.
L7 AI, Society and Rules	5 AI applications in society and industry; 4 AI policies and accountability	Analyse how AI is used across sectors and who is accountable. Explain how national policy and guidance shape responsible AI.	Sector table (what the AI does, what the person decides, a benefit, a risk, a rule) and a two-minute argument with the best reason from the other side.	Summative: Case study and structured debate	PartlyPolicy is taught from a short Local guidance box that summarises public UAE documents, not from the documents themselves, and the lesson is not yet checked with UAE educators.

Not covered yet. Cycle 3 (Grades 10 to 12): not covered. A4 reaches about Grade 9. Strand 5, Cycle 2: how AI is used across sectors, and global challenges such as sustainability and humanitarian aid, have no lesson. Strand 4: how AI is governed through national policies has no lesson. Strand 2, Grade 5: the main parts of a neural network have no lesson. Strand 6, Cycle 2: using generative AI to create content is limited to grown-up-led tests and Mavo. Assessment: the framework gives principles and example task types but no rubrics or performance levels, so teachers set the standard.

How Wondi complements the gaps

How Wondi complements the UAE gaps
Gap	The school covers it with	Wondi today	Planned, subject to funding
Cycle 3 (Grades 10 to 12)	The school's own AI literacy lessons and the Ministry's implementation guides.	A4 gives Grades 6 to 9 a base in models, bias and ethics, and a reasoned argument.	A Cycle 3 pathway with an ethical case study and a project portfolio.
AI across sectors and global challenges (strand 5)	Science, social studies and project lessons.	A1 L2 spots everyday AI, A2 L7 and A4 L7 look at AI at work across sectors, A3 L7 plans an AI-for-good helper, and A3 L5 judges real and made-up content.	An AI for good project for A3 and A4, on sustainability and humanitarian uses.
AI governance and national policy (strand 4)	The teacher, using the Ministry's guides and national policy documents.	A4 L6 covers privacy, copyright and a personal code of conduct. A3 L7 and A4 L7 add who is responsible and how national guidance shapes responsible AI.	A lesson on who makes the rules for AI, with a teacher guide to UAE policy.
Parts of a neural network (Grade 5)	The teacher, with the Ministry's examples.	A4 L1 to L3 teach prediction, training, testing and over-fitting.	A short unplugged lesson on the layers and parts of a network.
Creating content with generative AI (strand 6)	School-approved tools, under the Ministry's Safe and Responsible Use Guide for Grades 7 to 12.	Mavo's Prompt lab lets children aged 9 to 14 try prompts in a supervised chat that a parent can read.	Teacher notes on using Mavo within school rules.
Classifier tool for A2 to A4 (in development)	Not applicable.	Lesson plans are complete. Drawn example cards stand in.	The in-browser classifier, so pictures and sounds stay on the device.

UAE National AI Literacy Curriculum Framework (Ministry of Education) (opens in a new tab)

Gaps and sources

What Wondi does not cover. We mapped to the framework’s strands, not to each grade’s individual learning outcomes, which sit in the framework’s appendix and the Ministry’s implementation guides. Wondi does not cover Cycle 3 (Grades 10 to 12) or the parts of a neural network. Lesson 7 in each course now gives some coverage of AI across sectors, global challenges and national AI policy, but these are new, have not yet been classroom-tested or reviewed by UAE educators, and are not claimed as strong. The framework publishes assessment principles and example tasks but no rubrics, so teachers set the standard. Our lesson-by-lesson check is in uae-assessment.mjs.

Sources

India IndiaNEP 2020 and NCF-SE 2023, with a new Computational Thinking and AI curriculum for Classes 3 to 8

What it asks of children

The Ministry of Education announced in October 2025 that AI and Computational Thinking would start from Class 3 in the 2026 to 2027 school year, framed with CBSE, NCERT and state education departments. Reports say CBSE has since launched the curriculum, with teacher handbooks and support material on DIKSHA. Check CBSE and NCERT for the final materials.

Classes 3 to 5: computational thinking woven into subjects such as maths, languages and environmental studies, through puzzles, games and stories.
Classes 6 to 8: foundational AI concepts alongside computational thinking.
Themes: logical reasoning, problem solving, pattern recognition, ethical use of technology, digital literacy and critical thinking.

How Wondi lines up

Wondi course	Stage	Fit	What lines up
A1 Smart or Not?Explorers, ages 5 to 7	Foundational stage (before Class 3)	Light fit	Readiness only: the announced curriculum starts at Class 3. Sorting and pattern spotting (A1 L3) and safe, sensible use of technology (A1 L5, L6). Design a smart helper and say who checks it (A1 L7).
A2 Teach a MachineBuilders, ages 7 to 9	Classes 3 to 4	Partial fit	Pattern recognition: sort and train a picture sorter (A2 L2, L3). Logical reasoning and problem solving: test it, count errors, improve (A2 L4). Ethical use: fairness of examples (A2 L5). Ethical use: AI at work and who stays in charge (A2 L7).
A3 How Machines LearnCreators, ages 9 to 11	Classes 5 to 6	Partial fit	Pattern recognition and logical reasoning: build and test a classifier (A3 L2, L3). Critical thinking and digital literacy: real, fake or made-up content (A3 L5). Ethical and responsible use of AI (A3 L6). Ethical use: plan an AI helper for a good cause and say who is responsible (A3 L7).
A4 AI in PracticeCoders, ages 11 to 14	Classes 6 to 9	Strong fit	Foundational AI concepts: how models predict, training and testing (A4 L1 to L3). AI in everyday life: language models and made-up answers (A4 L5). Ethical use of technology: bias, privacy and copyright (A4 L4, L6). AI in everyday life: AI across sectors, accountability and a reasoned argument (A4 L7).

Gaps and sources

What Wondi does not cover. India’s new curriculum is being rolled out in stages and its module lists were not available to us, so we mapped to its published themes and class bands. Wondi is in English only for now.

Sources

US United States2026 CSTA PK–12 Computer Science Standards and the AI4K12 Five Big Ideas (voluntary; each state sets its own standards)

What it asks of children

The US has no national curriculum, so there is no single list to match. Two voluntary frameworks guide most AI-literacy work. The CSTA PK–12 Computer Science Standards were revised in 2026 and now tag AI-related standards across their five concept areas, with Machine Learning as a strand in every grade band. AI4K12 organises AI learning into Five Big Ideas: Perception, Representation and Reasoning, Learning, Natural Interaction and Societal Impact. Executive Order 14277 (April 2025) promotes AI literacy and K-12 resources but sets no standards. States often base their own computer science standards on CSTA’s, so check your state.

Grades PK to 2: recognise that AI systems use patterns in data to make decisions, tell what people do well from what computing technologies do well, and know that people design technology (CSTA EK-ALG-ML-02, E1-ALG-ML-02, E1-SOC-HU-15, E2-ALG-ML-02).
Grades 3 to 5: see how a machine learning model changes with its training data, train a model to classify, tell human learning from machine learning, and decide when to use technology (CSTA E3-ALG-ML-02 to E5-ALG-ML-02, E4-SOC-HU-18, E5-SOC-HU-18).
Grades 6 to 8: explain how a model makes predictions, improve its accuracy and reduce bias, evaluate its limits, weigh the risks of collecting personal data, and decide when AI is appropriate (CSTA MS-ALG-ML-06 to 08, MS-DAT-IM-28, MS-SOC-ET-40).

How Wondi lines up

Wondi course	Stage	Fit	What lines up
A1 Smart or Not?Explorers, ages 5 to 7	Grades K to 2 (CSTA PK–2)	Strong fit	Humans and machines: what people do well, what machines do well, and that people design them (A1 L1, L2). CSTA EK-SOC-HU-15, E1-SOC-HU-15. AI systems learn patterns from examples, unlike fixed-rule machines (A1 L2). CSTA E1-ALG-ML-02. Notice features, sort with a teaching pile and a testing pile (A1 L3). CSTA EK-ALG-ML-02, E2-ALG-ML-02; AI4K12 Learning. AI makes mistakes, private data stays private, and our rules for smart helpers (A1 L4 to L6). AI4K12 Societal Impact. Design a smart helper, name who checks it and who it might leave out (A1 L7). CSTA E1-SOC-HU-15; AI4K12 Societal Impact.
A2 Teach a MachineBuilders, ages 7 to 9	Grades 2 to 4 (CSTA E2 to E4)	Strong fit	Tell AI that learns from examples from technology that follows fixed rules (A2 L1). Examples teach machines: change the training set and the model changes (A2 L2, L3). CSTA E2-ALG-ML-02, E3-ALG-ML-02, E5-ALG-ML-02. Test on new examples and count right and wrong (A2 L4). AI4K12 Learning, grades 3 to 5. Unfair or incomplete examples give wrong or biased results (A2 L5). CSTA E2-DAT-IM-10, E4-ALG-ML-02; AI4K12 dataset bias. Present a model, how it was tested and a fairness note (A2 L6). Where AI helps at work and where a person decides (A2 L7). CSTA E4-SOC-HU-18; AI4K12 Societal Impact.
A3 How Machines LearnCreators, ages 9 to 11	Grades 4 to 6 (CSTA E4 to MS)	Partial fit	Choose between rules and learning for a problem (A3 L1). An early match for CSTA MS-ALG-PS-04. Training data and test data, then build and train a classifier (A3 L2, L3). CSTA E4-ALG-ML-02, E5-ALG-ML-02; AI4K12 Learning. Accuracy and the two kinds of mistake (A3 L4). AI4K12 Learning; an early match for CSTA MS-ALG-ML-08. Check whether content is real, fake or made-up before believing or sharing it (A3 L5). CSTA E3-SOC-ET-17, E5-SOC-ET-17. Decide when it is wise to use AI and when it is not (A3 L6). CSTA E5-SOC-HU-18. Plan an AI helper for a good cause and say who is responsible (A3 L7). CSTA E5-SOC-HU-18; AI4K12 Societal Impact.
A4 AI in PracticeCoders, ages 11 to 14	Grades 6 to 9 (CSTA Middle School)	Partial fit	Features, labels and predictions, then a nearest-neighbour classifier in Python (A4 L1, L2). CSTA MS-ALG-ML-06; AI4K12 Learning, grades 6 to 8. Training, testing and over-fitting (A4 L3). CSTA MS-ALG-ML-07, MS-ALG-ML-08. Bias in data and in decisions (A4 L4). CSTA MS-ALG-IM-10, MS-DAT-IM-29. Language models predict likely next words and can make things up (A4 L5). CSTA MS-SOC-ET-40. Ethics, privacy, copyright and a personal code for using AI (A4 L6). CSTA MS-DAT-IM-28, MS-SOC-HU-43. AI across sectors, who is accountable and a reasoned argument (A4 L7). CSTA MS-SOC-ET-40; AI4K12 Societal Impact.

Gaps and sources

What Wondi does not cover. Wondi does not teach AI4K12’s Perception (how computers see and hear), neural networks, reinforcement learning or decision trees; analysing AI-generated code (CSTA MS-PRO-RD-18) or cybersecurity (MS-SYS-SE-33); or the high-school AI specialty standards. Hands-on use of AI tools is limited to optional tests led by a grown-up. We read the CSTA standards and AI4K12’s Learning chart in full, but only the overview pages for the other four Big Ideas, so those ratings are provisional. We did not check any individual state’s standards, and grade levels are approximate.