Your school has an AI policy. Your teachers have attended demonstrations. Your students already know how to ask a chatbot for an answer. But what, exactly, have they learned about artificial intelligence—and what can they still do without it?
That distinction should shape the next school leadership conference. On 18 June 2026, the OECD and European Commission released Empowering Learners for the Age of AI, a framework for primary and secondary education. It describes 19 competences across four domains: Engage with AI, Create with AI, Manage AI, and Shape AI. It offers learner expectations and classroom scenarios, not merely a list of recommended tools. OECD publication · European Commission announcement.
For a superintendent, principal, school network leader or conference organiser, the real opportunity is to turn this shared vocabulary into a decision: Which kinds of learning should our school protect, redesign and evaluate in an AI-rich world?
The shift: from AI tools to AI literacy
An AI demonstration answers, “What can this application do?” AI literacy asks, “What can a learner understand, question, create, decide and take responsibility for?”
The 2026 framework's four domains help make that difference tangible:
Engage with AI: encounter and interpret AI-generated information and recommendations with judgement.
Create with AI: collaborate with AI while preserving purpose, authorship and human creativity.
Manage AI: make intentional decisions about when, why and how AI should support work.
Shape AI: understand and question the systems, values and choices that determine what AI does in society.
This is a useful organising lens, not an internationally compulsory curriculum or a ready-made certification. Its authors explicitly encourage adaptation to local contexts. Schools must still consider age, existing curriculum, disability access, infrastructure, family expectations and applicable national rules. AILit Framework.
A strong leadership conference should therefore achieve more than collective enthusiasm or collective anxiety. It should help people make difficult, concrete choices together.
Question 1 — Are we educating AI users, or independent thinkers?
Imagine two students who submit polished work. One can explain the choices, identify a weak argument, and reproduce essential reasoning independently. The other cannot. Their outputs look similar; their learning is not.
Start the conference with this distinction. Ask leaders which abilities students must demonstrate before, during and after AI-supported work. For example, before asking for feedback from a model, a student might sketch an argument. During the interaction, they might challenge a doubtful claim. Afterwards, they might explain the final reasoning without the tool.
This is not an argument for forbidding AI. It is an argument for making the learner's thinking visible. The OECD–EC framework emphasises critical evaluation and ethical, creative participation; it does not equate frequent tool use with literacy.
Conference output: an agreed definition of independent learning in AI-supported tasks, including two or three observable student behaviours.
Question 2 — Which AI competences belong in which subjects and age groups?
A framework can become another poster on the staffroom wall unless educators connect it to real curricula.
A primary teacher might help pupils recognise that a digital recommendation can be generated by a system rather than a person. A secondary history teacher might ask students to compare an AI summary with original sources. An upper-secondary science class might consider the data and assumptions behind an AI-assisted prediction.
These are illustrative teaching possibilities, not claims that the framework mandates these precise lessons. The point is progression: the same underlying competence can take different forms at different developmental stages.
An effective conference makes space for subject teams to choose one competence, map it against existing learning objectives, and design one developmentally appropriate task. Teachers should not have to rebuild every course at once.
Conference output: a cross-curricular pilot map covering grades, learning objectives, AI-literacy domains and responsible staff owners.
Question 3 — What should students know about reliability, bias and evidence?
AI can produce confident answers that are incomplete, inappropriate or wrong. The educational response should not be merely “always double-check”; learners need a method for checking.
Give conference participants a sample answer to a realistic curriculum question, then ask them to identify which claims need supporting evidence, where primary sources can be found, and whether the model's framing leaves out important perspectives. Compare what a younger learner and an older learner could reasonably be expected to verify.
The key distinction is between accuracy checking and epistemic judgement: deciding which evidence deserves trust, who has relevant expertise, and when uncertainty should remain visible.
UNESCO's 2024 student framework likewise foregrounds a human-centred mindset, AI ethics, techniques and applications, and system design. It is a complementary reference, not an identical document. UNESCO AI Competency Framework for Students.
Conference output: a usable evidence-checking routine that students can practise and teachers can assess.
Question 4 — How will assessment reveal learning rather than polished production?
If students can generate a report, presentation or plausible explanation quickly, the final product alone tells teachers less about what the learner knows.
That does not make authentic projects obsolete. It changes the evidence we need. Schools can combine a well-designed product with a short oral explanation, a process record, an in-class reasoning task or a reflection on why specific AI suggestions were accepted or rejected.
The right combination depends on age, subject, accessibility and assessment purpose. An oral defence, for example, is not a universal solution: learners with language, communication or anxiety-related barriers may need equivalent accessible ways to demonstrate competence.
A leadership conference should debate what valid evidence of learning looks like when AI assistance is permitted, disclosed or deliberately excluded. This requires fairness as well as academic integrity.
Conference output: one revised assessment task per department, with transparent AI-use expectations and an explanation of what it actually measures.
Question 5 — Are teachers being developed as professionals, or simply trained as tool operators?
A school can buy licences overnight. Professional judgement and shared teaching practice develop more slowly.
UNESCO's AI Competency Framework for Teachers, published in 2024, sets out 15 competences across five dimensions, including human-centred values, ethics, foundations and applications, AI pedagogy, and professional learning. It recognises teachers as decision-makers, not passive consumers of software. UNESCO teacher framework.
Use the conference to clarify what teachers must understand, what they should experiment with, and where they retain authority to decline an AI-supported approach. A mathematics teacher may need help designing productive hints; a safeguarding lead may need a discussion of data boundaries; a school principal may need an approach for handling parental questions.
Staff development should include time for planning, peer review, practice and follow-up. One inspirational speech can open a conversation; it cannot substitute for the sustained support needed to change routines.
Conference output: a differentiated professional-learning plan with practical follow-up, not just attendance certificates.
Question 6 — Who decides the boundaries for safety, privacy and equity?
AI literacy is not a technology department's responsibility alone.
A school may have to weigh useful functionality against the collection of student data; efficiency against teacher workload; personalised feedback against unequal access; creativity against opaque intellectual-property expectations. Leaders should engage teachers, students, families, safeguarding staff and relevant legal or data-protection expertise.
A conference can use carefully designed scenarios instead of vague principles. What happens when a teacher wants to paste a student's sensitive work into an external service? When a family objects to a particular tool? When an AI tool offers different quality of feedback across languages?
National policies vary, so the OECD–EC framework should not be mistaken for legal compliance advice. Rather, its competencies provide an educational starting point for discussing governance in a local context.
Conference output: a named decision owner, a short risk-escalation process and a list of unresolved policy questions to assess with qualified specialists.
Question 7 — What will be different 90 days after the conference?
Perhaps the most important question is also the least glamorous. Many events end with a compelling keynote and an overflowing list of ideas. Then timetables, budgets and existing habits return.
Before the conference, ask every participating team to bring one real learning problem. At the event, use the OECD–EC domains to examine it. Before participants leave, agree a small pilot, its owner, the student-learning evidence to collect, and a review date.
Success should not be measured by how many staff tried a chatbot. Depending on the pilot, more meaningful observations might include whether students can explain reasoning, identify weaknesses in AI claims, make informed choices about when to use a tool, or transfer a skill to a new task. Evidence should be interpreted cautiously: a short pilot does not prove long-term impact.
Conference output: a 30–60–90-day action plan with specific responsibilities, review points and a way to hear student and teacher feedback.
Turn the seven questions into a practical leadership agenda
Here is an illustrative one-day format that organisers can adapt; it is not a report of an event Learning Coach has already delivered.
| Session | Leadership question | Takeaway |
|---|---|---|
| Opening keynote | What does AI literacy mean beyond tools? | Shared purpose and language |
| Curriculum lab | Where do the four domains fit? | Grade/subject pilot map |
| Assessment clinic | How will we see student thinking? | Revised task and evidence criteria |
| Governance roundtable | Who owns boundaries and exceptions? | Risk and decision map |
| Action-planning session | What changes in 90 days? | Named pilots, owners and review dates |
A keynote is particularly useful when an organisation needs common language and a reason to change. Facilitated workshops are more suitable when teams must redesign tasks, resolve local trade-offs or practise new routines. Many school networks will benefit from both, but not every event requires a full-day format.
For organisers, ask prospective speakers and facilitators to demonstrate how they distinguish scientific evidence from aspiration, adapt to the age range, address ethical tensions, and leave participants with decisions rather than slogans.
The Learning Coach perspective: AI should extend learning, not replace it
The distinctive question for Learning Coach is not whether AI will enter school life. It already has. The question is who remains the active learner.
From this perspective, the most valuable use of a framework is to protect learner agency: the ability to set goals, exercise judgement, recognise gaps, seek help intelligently and retain responsibility for decisions. A student's fluency with prompts is not necessarily fluency with ideas. And a school's technology maturity is not the same as its learning maturity.
This is an editorial argument, not a claim that the OECD framework has independently validated a particular Learning Coach method. It can be tested through the quality of classroom tasks, student explanations, teacher decisions and the questions leaders are prepared to ask.
That is also why a school leadership conference matters. Teachers cannot carry this transition alone. The organisation must allow time for professional adaptation, make policy understandable, acknowledge resistance as a source of information, and choose meaningful educational outcomes over performative adoption.
Frequently asked questions
Is the OECD's 2026 AI Literacy Framework mandatory for schools?
No. The OECD and European Commission describe it as a common reference for primary and secondary education, intended to support adaptation and implementation. Schools must consult their own curriculum requirements and applicable regulations rather than assume the framework itself creates a legal obligation.
Does AI literacy mean every student must use generative AI?
No. Understanding AI includes evaluating systems, their limitations, and their social consequences. Some learning experiences can examine AI critically without requiring every pupil to have an account or use a generative tool directly.
Can a single school keynote deliver the framework?
A keynote can introduce the framework, challenge assumptions and align decision-makers. Meaningful implementation also needs curriculum design, teacher development, governance, assessment choices and follow-up. No one event can guarantee better learning outcomes.
How should an international school adapt the framework?
Begin with the age range, languages of instruction, local safeguarding and privacy expectations, existing curricula and student access. Map a limited set of competencies to real subject tasks and evaluate what students learn, not just which tools they use.
The conversation worth bringing to your next event
The 2026 AI Literacy Framework gives leaders a more useful starting point than “Which AI tools should we buy?” It asks what young people need to understand, make, manage and shape in a society influenced by AI.
The seven questions above turn that ambition into a conference agenda that can produce something visible: a curriculum pilot, a better assessment, a clearer teacher-learning plan and accountable decisions.
A future-ready school is not the school that uses AI most often. It is the school that can explain what its students are learning because—and sometimes despite—the tools around them.
Planning a school leadership summit, professional development day or education conference on responsible AI and independent learning? Explore Learning Coach's international keynotes and request a proposal.
References
OECD & European Commission (2026), Empowering Learners for the Age of AI: An AI Literacy Framework for Primary and Secondary Education. Published 18 June 2026.
European Commission (2026), New AI Literacy Framework helps schools prepare learners for the age of artificial intelligence. Published 18 June 2026.
AILit Framework official project website. Consulted 26 October 2026.
Editorial note: illustrative scenarios and the suggested conference agenda are examples, not reports of events delivered or independently tested.
