The Five Refined AI Design Principles Explained in Simple Language, with Classroom Examples from Early Years to Grade 5 and a Practical Starting Guide for PYP Educators
The IB has shared an update on its AI Design Principles. In short, these are five principles that guide how AI should be used across the whole IB community, not only in PYP, shaped with feedback from educators around the world.
If you have any doubts or confusion about how to bring this update into your PYP classroom in a practical way, from Early Years to upper primary, this article is a guide for all PYP teachers. We have added grade-wise examples and easy tables so that you can understand each idea and use it with confidence.
About this guide
An independent guide by Inquiry Classroom, based on the IB’s March 2026 and 22 September 2026 updates. The principle rows are our own summaries, not direct quotations, and the examples, checklists and plans are our suggestions, not IB requirements. Follow your school’s AI policy. Information current as of 11 October 2026. Full principles: IB’s September update.
1. Quick summary: what you need to know
| Question | Short answer |
|---|---|
| What is it? | Five design principles that guide how AI should be used across the IB community. |
| Is it final? | No. The principles are moving into final review and will keep evolving. The IB says it will review community feedback each quarter. |
| Is it a rule I must follow tomorrow? | The IB pages we reviewed announce no new PYP requirement. They describe principles and tools being built. Your school policy still comes first. |
| Is it only for PYP? | No. It covers the whole IB community, so the wording is general. This guide shows how it applies in PYP classrooms. |
| What changed since March? | Two principles were renamed and reshaped after feedback: the first now stresses protection, and the educator one now stresses shared responsibility. The wording of all five descriptions was refined as well. |
| What should I do now? | Learn your school policy, try small teacher-led uses, keep children thinking, protect data, and keep notes on what works. |
2. What happened between March and September 2026
| When | What happened | What it means for you |
|---|---|---|
| March to April 2026 | The IB shared draft AI Design Principles and asked for feedback. Co-creation sessions followed at the IB Global Conference in Mumbai (March) and the African Education Festival in Johannesburg (April). | It was a draft to discuss, not a rulebook. Teachers and school leaders helped shape the changes. |
| 18 July 2026 | A plenary at the IB Global Conference in Denver asked what role each person plays in the AI learning ecosystem. | The focus moved towards shared, values-led responsibility. |
| 22 September 2026 | The IB published its update, “Co-creating our approach to AI”. Feedback came from more than 1,500 educators in 119 countries, through an online survey and 11 co-creation sessions. | Principles refined. Still not final. |
| 22 to 24 October 2026 (upcoming) | As of the 22 September update, the IB planned to share progress at the IB Global Conference in Lyon and in virtual workshops (co-creation sessions run between August and November). | Check the IB site for what has been shared since. The IB said it would share progress, not that final guidance will be released there. |
| Later | In March, the IB indicated that it would release draft red lines and a practical toolkit alongside the final principles later in the year. | No confirmed publication date is available yet. Treat the draft guidance as provisional and wait for the official release before considering any toolkit items final. |
What changed between the draft and the update
| Principle | March draft | September update | Why it changed (IB’s feedback) |
|---|---|---|---|
| Principle 1 | Caring and balanced: AI for human flourishing | Caring and protective: AI for human and planetary flourishing | Feedback highlighted the need to strengthen learner protection and extend the principle to include planetary well-being. |
| Principle 3 | Educator agency: guided by capability, grounded in responsibility | Shared responsibility: across the IB ecosystem | The draft put too much on educators and too little on organisations and tool makers. One educator said safeguarding cannot sit with students or educators alone. |
| Principles 2, 4, 5 | Inquiry-driven; Safe and transparent: AI that is accountable; Continuously adapting | The titles remain unchanged. Principle 4’s subtitle is now “Accountable use of AI.” | The descriptions were refined to clarify expectations. Principle 2 emphasises preserving opportunities for learners to demonstrate their own growth, while Principle 5 highlights the need to discontinue AI use when it no longer benefits learners. |
Teacher takeaway
For educators, this may be reassuring: the IB emphasises shared responsibility rather than placing the burden on teachers alone.
3. Common doubts, answered
| Your doubt | Clear answer |
|---|---|
| Do I have to use AI in my lessons? | The updated principles do not require AI to be used in every lesson. They focus on using AI thoughtfully when it adds educational value. Discussion, hands-on experiences, reading and human interaction may be more appropriate in many learning situations. |
| Is this the final version? | Not yet. The IB describes the updated principles as moving into final review and indicates that they will continue to evolve. Avoid making major changes to unit planners based solely on wording that may still change. |
| Do students need to use AI? | The update does not mandate student use of AI. Where student use is permitted by school policy, the principles can guide responsible and purposeful integration. |
| Should I ban AI in my classroom? | The IB encourages schools to take a thoughtful approach to AI rather than relying only on a blanket ban. Teachers should receive clear guidance and support to make informed decisions about AI use in the classroom. As a PYP teacher, follow your school’s AI policy and consult your PYP Coordinator or school leadership whenever clarification is needed. |
| Can children from Early Years to Grade 5 use AI tools themselves? | The IB pages give no age rules. This depends on your school policy and each tool’s own terms, and many general chatbots limit younger users. Age limits, consent requirements and tool terms also vary by country and provider, and school policy comes first. Our suggested starting point, not an IB rule: from Early Years to Grade 3, keep AI teacher-led (the teacher shows, the children discuss). For Grades 4 and 5, student use only with an approved tool and clear supervision. |
| Which AI tools are approved? | The September update does not provide a definitive list of approved AI tools. Schools must establish their own approval processes. Consult your PYP Coordinator or school leadership, and check each tool’s age restrictions, terms of use and data-protection arrangements. AI tool selection and vetting is among the draft toolkit resources still in development. |
| What does “ecosystem” mean? | In this context, the AI ecosystem includes the different stakeholders involved in educational AI: the IB, schools, teachers, students, families and the organisations that develop AI tools. |
| What are “red lines”? | The IB’s March communication described red lines as proposed, non-negotiable boundaries for AI use. One illustrative draft example was that high-stakes decisions involving grading, student progression and well-being should not be fully automated. As a complete list was not available in the pages reviewed, treat this as developing guidance and continue to follow school policies and applicable regulations. |
| What is the “toolkit”? | The toolkit refers to practical resources being co-designed with the IB community to support responsible AI implementation. The September update identifies nine draft tools, including an AI policy and governance template, AI literacy and capability building, AI tool selection and vetting, assessment and academic integrity, and parent and community conversations. Two additions are AI councils and professional development. These resources remain in development and should not be treated as mandatory requirements. The first priorities identified are the policy and governance template, AI literacy for educators, and student AI leadership councils. |
| What does “planetary flourishing” mean? |
It refers to the well-being of both people and the planet. AI systems require electricity to operate, and data centres may use water to cool the computers that process AI requests. These activities can have an environmental impact. |
| Does the update change PYP assessment or academic integrity rules? | The pages reviewed do not establish new assessment or academic integrity requirements. Although an assessment and academic integrity resource appears in the draft toolkit, it remains in development. Continue following the IB’s existing academic integrity guidance and your school’s policies. If your school has no clear guidance on acknowledging AI assistance, seek clarification from your PYP Coordinator or school leadership. |
| Am I responsible if a child misuses AI? | he update presents responsibility for educational AI as shared across the IB, schools, families and technology providers. Teachers remain responsible for following school policies, modelling responsible use, protecting learners and seeking guidance when uncertain. |
4. The five principles, with examples from Early Years to Grade 5
For each principle you will find what it means, what to do and avoid, and one example for every stage from Early Years to Grade 5. Use only school-approved tools and check any AI content yourself before showing it to children.
Principle 1: Caring and protective TITLE CHANGED IN SEPTEMBER (WAS: CARING AND BALANCED)
AI for human and planetary flourishing
AI for human and planetary flourishing
| In one line | AI use should protect children’s wellbeing, privacy and meaningful human relationships. |
|---|---|
| The IB’s refined principle, in our words | AI in an IB education should safeguard learners’ emotional, social and cognitive development, and support learning, inclusion and access. It should protect the human relationships and trust at the heart of an IB education, work across the many contexts IB schools operate in, and reflect a commitment to planetary flourishing by weighing AI’s environmental cost and using it where it adds educational value. |
| In simple words | Before using any AI, ask three things: Is it good for the child? Is it fair to every child? Does it replace a real human connection? If AI risks replacing a meaningful human connection, reconsider the use and choose a more appropriate approach. |
| Do |
|
| Avoid |
|
| Early Years example | Theme: Sharing the Planet. In circle time the teacher shows one AI-made picture of a playground and asks, “Who can play here? Who is missing?” Children answer through talk and drawing. Teacher-led only. |
| Grade 1 example | Theme: Who We Are. The teacher reads a short AI-written story about a kind friend, checked first. Children say what real friends do that a computer cannot, like listening and hugging. |
| Grade 2 example | The teacher presents two AI-generated pictures of a happy neighbourhood. Learners identify who is represented, notice who may be missing, and draw additional people to make the neighbourhood more inclusive. |
| Grade 3 example | Theme: Who We Are. The class reads an AI-written paragraph about friendship, compares it with how real friends behave and decides what AI cannot do. |
| Grade 4 example | Theme: Sharing the Planet. Using sources chosen by the teacher, learners find out that AI needs electricity and water. They discuss when using AI is worth it and when a book or a chat would do. |
| Grade 5 example | Learners read an AI-written story about a community and check it for stereotypes. They rewrite it more fairly and explain what they changed. |
| Ask yourself | Could this use of AI exclude, misrepresent or disadvantage any child? |
Principle 2: Inquiry-driven
AI that deepens learning
AI that deepens learning
| In one line | AI should make children think more, not less. |
|---|---|
| The IB’s refined principle, in our words | AI should strengthen inquiry, critical thinking, creativity and collaborative problem solving, and help learners become active explorers of their own education. It should serve the learner’s own growth, whether intellectual, ethical, social or emotional, and that growth belongs to the learner. Learning and assessment should keep the moments where growth is visible and recognisably the student’s own. |
| In simple words | If AI gives the answer and the child copies it, that is not inquiry. If AI gives an idea and the child tests it, questions it or improves it, that is inquiry. |
| Do |
|
| Avoid |
|
| Early Years example | Theme: How the World Works. Children guess what a plant needs to grow. They put two seedlings, or already sprouted beans, in the light and in a dark cupboard, water both and look every day. Later the teacher says, “A computer told me plants need light and water to grow. Was it right?” |
| Grade 1 example | Children watch their shadow outside in the morning and the afternoon and say what they noticed. Then the teacher reads what AI says about shadows and the class checks it against what they saw. |
| Grade 2 example | Children predict which objects will float, then test them in a water tub. The teacher shows what AI said and the class sorts which answers matched the test. |
| Grade 3 example | The teacher asks AI for three questions about the water cycle and shows them. Learners choose the best one, explain why, and add two of their own. |
| Grade 4 example | Learners see an AI explanation for why a toy car travels farther on a smooth floor. They run fair tests and decide which explanation the evidence supports. |
| Grade 5 example | Learners collect possible reasons for wilting school-garden plants, including AI-suggested ones. They design a fair test for the most likely reason and report results. |
| Ask yourself | Is AI making this child think more or less? |
Principle 3: Shared responsibility TITLE CHANGED IN SEPTEMBER (WAS: EDUCATOR AGENCY)
Across the IB ecosystem
Across the IB ecosystem
| In one line | Nobody carries AI safety alone: not the teacher, not the child, not the parent. |
|---|---|
| The IB’s refined principle, in our words | Educators shape how AI is used in learning, but they cannot do it alone. Responsible AI use is the work of the whole IB community: schools, students, parents, technology providers and the IB itself. The IB states that “Banning AI outright is not a plan,” and that neither is leaving every teacher to figure it out alone. The IB has an obligation to set direction, give practical guidance and support the community, and the community shapes that direction through its practice, judgement and feedback. |
| In simple words | You are not the AI police. The IB, schools, families and tool makers all share the job. Your part is to follow school policy, show good habits and ask when you are unsure. |
| Do |
|
| Avoid |
|
| Early Years example | With the teacher, children agree two picture rules: “We ask a grown-up first” and “We keep our names private.” The rules are shown as drawings and repeated in circle time. |
| Grade 1 example | Each child draws and tells one safe-use rule. The teacher puts them together into a class rule chart with pictures and simple words. |
| Grade 2 example | Children role-play “ask a grown-up first” situations, then explain the rules to a younger buddy class. |
| Grade 3 example | Learners and teacher make an “Our AI Helper Rules” poster. A short note goes home so families hear the same message. |
| Grade 4 example | Learners interview a family member or another teacher about how they use AI and what rules they follow, then compare answers in class. |
| Grade 5 example | Learners draft a class agreement covering when AI is allowed, how to say AI helped and how to check facts. The class reviews it at the end of the unit. |
| Ask yourself | Does every child, parent and teacher in my class know what the AI rules are? |
Principle 4: Safe and transparent
Accountable use of AI
Accountable use of AI
| In one line | Protect data, know what the tool does and where it fails, and keep humans in charge of big decisions. |
|---|---|
| The IB’s refined principle, in our words | Every learner’s data rights, privacy and safety are non-negotiable. AI use in the IB community should meet strong safeguarding, transparency and data protection standards. Accountability always rests with people, with the IB, with schools and with providers, never with the AI systems themselves. In the IB’s words, “AI may inform, but accountable humans must decide.” |
| In simple words | Never put identifiable student information into an unapproved tool. Check AI facts before sharing them. AI may inform, but accountable people must make the decisions, especially high-stakes ones about a child’s grades, progression or wellbeing. |
| Do |
|
| Avoid |
|
| Early Years example | The teacher shares a funny wrong AI answer prepared in advance, such as “a cat has six legs”. Children count the legs on a toy cat and say, “Is that right?” |
| Grade 1 example | The teacher reads aloud an AI sentence about an animal that contains one mistake the teacher has already found. Children check it against a picture book. |
| Grade 2 example | Children get two AI sentences about an animal and use True and False cards after checking classroom books. |
| Grade 3 example | Learners check one AI paragraph about animals against two books and underline what matches and what does not. |
| Grade 4 example | Learners keep an “AI mistake log”. They record any error found, the correct fact and the source that proved it. |
| Grade 5 example | Under each piece of work, learners add a short note: what AI helped with, what they checked, and what they still are not sure about. |
| Ask yourself | Can the learner explain where information came from and how it was checked? |
Principle 5: Continuously adapting
Evidence-led, always improving
Evidence-led, always improving
| In one line | Try small, check what happened, then keep, change or drop it. |
|---|---|
| The IB’s refined principle, in our words | AI should prove its educational value through evidence, reflection and honest evaluation. Responsible engagement means learning from practice and listening to students, parents, educators and schools. It means being willing to change course, or stop, when a use no longer serves learning, wellbeing, equity, safety, sustainability or trust. The IB says both inaction and recklessness fail learners, and calls this “mindful innovation”. |
| In simple words | None of this is permanent. Treat AI use like any other strategy: test it, look at the evidence, adjust. Choosing not to use AI can be appropriate when it adds no meaningful learning value, but ignoring the question altogether is not a plan either. |
| Do |
|
| Avoid |
|
| Early Years example | After the activity, children show a happy, middle or sad face to say if it helped them learn. The teacher notes a few comments. |
| Grade 1 example | After the activity, the teacher asks “Did it help us learn?” Children show thumbs up, middle or down. The teacher notes the count. |
| Grade 2 example | Learners draw one thing they learned from an AI-generated response and one thing the AI got wrong or left out. The teacher reviews their drawings to identify patterns in learners’ understanding and misconceptions. |
| Grade 3 example | Exit ticket: “What did we learn that AI did not tell us?” The teacher reads them and notes patterns. |
| Grade 4 example | Two groups research the same topic, one with AI support from the teacher and one without. The class compares the questions asked and the understanding shown. |
| Grade 5 example | Learners write a short reflection on how AI changed their thinking. The teacher compares it with earlier work and decides what to change in the next unit. |
| Ask yourself | What evidence shows this actually improved learning? |
5. Classroom quick guides
AI across the grades at a glance (our suggestions)
A quick planning view from Early Years to Grade 5.
| Focus area | Early Years | Grade 1 | Grade 2 | Grade 3 | Grade 4 | Grade 5 |
|---|---|---|---|---|---|---|
| Who uses the tool | Teacher only | Teacher only | Teacher only | Teacher mostly, learners observe | Teacher-led, learners may try with an approved tool | Learners may use with an approved tool and supervision |
| Main skill | Notice and wonder | Ask “Is that true?” | Check with books and tests | Compare AI with own ideas and sources | Find errors and give evidence | Evaluate, decide and reflect |
| Best AI role | Something to talk about | Something to question | A source of ideas to test | A source of questions | A source of claims to test | A thinking partner that is checked |
| Thinking shown through | Talk, play and drawing | Talk and drawing | Drawing, sorting, short sentences | Short oral and written explanations | Mistake log and fair tests | Reflections and process notes |
| Watch out for | Showing unchecked content | Showing unchecked content | Believing it without testing | Copying AI wording | Believing AI over evidence | Over-reliance, missing acknowledgement |
Traffic light guide: green, amber, red
Use this when you are not sure whether something is a good idea. It is our own guide, not the IB’s red lines.
| Level | What it looks like | Examples |
|---|---|---|
| GREEN: go ahead | Teacher-controlled, no student data, learners stay in charge of thinking. |
|
| AMBER: ask first | Involves children directly, or you are unsure about the tool. |
|
| RED: avoid | Risks privacy, replaces thinking or hands decisions to AI. |
|
6. A full Grade 5 example: one unit, step by step
Theme: Sharing the Planet
Possible central idea: Human choices influence the sustainability of natural environments.
Focus: Learners study school waste and propose an action.
| Stage | What learners do | What you do | Principle |
|---|---|---|---|
| 1. Tune in | Observe class waste. Write their own questions first. | Collect questions before bringing in any AI. | Inquiry-driven |
| 2. Explore | Look at waste-reduction ideas suggested by an approved AI tool. | Model saying “Interesting, but how do we know?” | Safe and transparent |
| 3. Investigate | Compare ideas with real school data, books and reliable sources. | Support evidence checking and discuss what is realistic. | Inquiry-driven |
| 4. Take action | Choose one idea and design a campaign or proposal for the school. | Make sure the choice is learner-led and fits the school. | Shared responsibility |
| 5. Reflect | Explain what AI helped with, what they checked and what they changed. | Assess reasoning, research and reflection, not only the final product. | Continuously adapting |
Questions to ask during the unit:
- Which of these ideas has evidence behind it?
- What do we need to find out before choosing?
- Would this work in our school? Why or why not?
- How can we tell if the AI information is correct?
- Which decisions should stay with us?
7. Our suggested four-week start plan for new teachers
You do not need to overhaul anything. Small and careful beats fast and risky.
| Week | Action | Done when |
|---|---|---|
| Week 1 | Read your school’s AI and academic integrity policy. Ask your coordinator the four questions below. | You know what is allowed in your school. |
| Week 2 | Try one teacher-only use, like drafting question starters for a unit. Check and edit everything. | You have used AI once safely. |
| Week 3 | Run one small, teacher-led learner activity from Section 4 for your grade. | You have watched how children respond. |
| Week 4 | Look at the evidence: what children said and produced. Decide keep, change or drop. Share with your team. | You have written down one clear next step. |
Four questions to ask your coordinator
- Which AI tools are approved for teachers and for students?
- Which grades may use AI directly, if any?
- How should learners acknowledge AI help?
- What information must never be entered into any AI tool?
8. For your PYP team: checklists, questions and jargon decoder
Checklists for teachers, school leaders and learners
| For teachers | For PYP coordinators and school leaders | For learners |
|---|---|---|
|
|
|
Five questions for your next team meeting
- Where could AI genuinely help our current Programme of Inquiry, and where would it not?
- What should children still do and explain without AI?
- What are our shared rules for data, accuracy and acknowledgement?
- How will we know if something is helping learning?
- Which questions should we send to our coordinator or the IB?
Jargon decoder: IB words in classroom language
| Word in the IB update | What it means in classroom language |
|---|---|
| Ecosystem | Everyone involved: IB, schools, teachers, learners, families, tool makers. |
| Governance | The rules, policies and decision-making around AI in a school. |
| AI literacy | Knowing what AI can do, where it fails and how to use it wisely. |
| Human oversight | A real person checks and decides, not the AI. |
| High-stakes decision | Something that affects a child’s future: grades, progression, wellbeing. |
| Red lines | Proposed non-negotiable limits on AI use. The full list is still to be published. |
| Co-creation | The IB shaping its approach together with the community. |
| Toolkit | Practical templates and guides the IB is building for schools. |
9. What to watch next
- Final review of the principles. The IB says they will keep evolving, with community feedback reviewed each quarter.
- The IB Global Conference in Lyon, 22 to 24 October 2026. As of its 22 September update, the IB planned to share progress there and in virtual workshops. Check the IB site for any newer news.
- The toolkit. Still in development. The first priorities are listed in Section 3.
- The red lines. Proposed in March, with the full draft list promised alongside the final principles. Watch for it.
- An IB white paper on its AI leadership role. The IB says it is being finalised.
Until the IB publishes final guidance, treat today’s principles as direction, not as a checklist to tick.
Conclusion
In short, the IB’s message is this: keep children thinking, keep their data safe, keep humans in charge, share the responsibility and keep checking what works. AI is not required in every lesson. A good PYP lesson with no AI at all still fits the principles.
If you remember one thing: start with the learning goal, not the tool. Then ask whether AI helps children think more.
At Inquiry Classroom we will update this guide when the IB publishes new guidance.
Official IB sources
- Shaping our approach to AI as a community (IB, March 2026)
- Co-creating our approach to AI: an update and learnings so far (IB, 22 September 2026)
Editorial note: This is an independent educational interpretation by Inquiry Classroom, based on the IB’s published pages. The IB’s AI Design Principles apply across the whole IB community, not only PYP, so this guide translates them for PYP classrooms. The classroom examples, traffic light guide, start plan, grade table, checklists and discussion questions are our suggestions, not IB-prescribed practice or policy. Check the IB’s official pages and your school’s policy before making decisions.
