What Belongs in Our Learning Toolkit in the Age of AI
As students and teachers return to classrooms for the 2026–27 school year, the usual back-to-school toolkit looks a little different. Alongside iPads, laptops, lesson plans and freshly sharpened pencils sits something far more powerful and far more complicated: artificial intelligence (AI).
AI is no longer something happening somewhere in the future. It is already helping students explain difficult concepts, generate ideas, translate languages, analyse information and receive feedback. Teachers are using it to develop resources, differentiate learning, generate questions and reduce some of the administrative load that competes with the human work of teaching.
The question for this school year, therefore, is no longer: “Should we use AI?”
A much more useful question is: “How do we use AI without giving away the very thinking that education is supposed to develop?” That distinction matters.
The OECD’s Digital Education Outlook 2026 makes an important observation: performing better on a task with generative AI does not necessarily mean that a student has learned more. AI can improve the product while, if used indiscriminately, reducing the cognitive effort involved in creating it.
So, as we begin another school year, perhaps our most important supplies are not technological at all. Hence, here are six things I believe every student and every teacher should have in their 2026 learning and teaching toolkit.
1. Curiosity before convenience
AI is extraordinarily good at giving us answers. Hence education must remain extraordinarily good at helping us ask questions.
The temptation this year will be to start every task with a prompt: Explain this. Summarise this. Write this. Solve this.
Instead, start with: What do I think? What do I notice? What don’t I understand? What would I like to know?
Then bring AI into the conversation. Ask it to challenge your hypothesis. Ask for another perspective. Ask it to identify weaknesses in your reasoning. Ask it to generate questions you have not considered. The goal should not be to become better at getting AI to think for us, but to become better at using AI to extend our own thinking.
That requires and nurtures curiosity.
2. A healthy dose of scepticism
AI can sound remarkably confident while being remarkably wrong. That makes critical thinking, source evaluation and verification more important, not less.
Students need to develop the habit of asking: Where did this information come from? Can I verify it somewhere else? Whose perspective is represented Whose perspective might be absent? Is this fact, interpretation or fabrication?
Teachers need the same habits. AI literacy is increasingly being understood not simply as knowing how to operate an AI tool, but as being able to understand, critically evaluate and use AI responsibly. The OECD and European Commission’s 2026 AI Literacy Framework places knowledge, skills and attitudes together at the heart of preparing young people for an AI-rich world.
Perhaps one of the most valuable phrases a learner can use this year is: “I am not sure. Let me check.”
3. The ability to know what NOT to outsource
This may be one of the defining learning skills of our time. Humans have always offloaded cognitive tasks. We use calculators instead of doing long division. Calendars remember appointments. GPS remembers routes. AI dramatically expands what we can offload. But can and should are not the same thing.
If I ask AI to improve the grammar in something I have already written, it may allow me to concentrate on my argument. If I ask AI to write the argument before I have thought about it, I may have outsourced the learning itself.
This year, students and teachers need to become better at identifying the difference between productive assistance and cognitive replacement. Cognitive offloading over cognitive outsourcing.
Before using AI, ask: What is the learning here?
If the struggle is part of the learning, don’t immediately remove the struggle. The indiscriminate cognitive offloading to AI is actually cognitive outsourcing which may weaken deep reading, sustained attention, perseverance and what it describes as cognitive stamina.
Sometimes the most educationally valuable response to a difficult task is still: Stay with it a little longer.
4. Transparency
Schools will continue debating AI detection, academic integrity and what constitutes acceptable assistance. But perhaps we should spend just as much time developing a culture of transparency.
Students should become comfortable saying: “I used AI to brainstorm possible approaches.” “AI helped me identify weaknesses in my first draft.” “I asked AI to explain this concept, then checked the explanation against these sources.”
Teachers should be equally transparent: “I used AI to help generate these examples, and I reviewed them.” “AI helped me adapt this resource for different reading levels.”
Instead of creating a perpetual game of catch the AI user, we can teach young people to document and explain how, why and when they used AI. Academic integrity in 2026 cannot simply mean proving that AI was absent. Increasingly, it must mean demonstrating that human thinking was present.
5. Human skills
There is a fascinating paradox at the centre of the AI revolution. The more capable machines become, the more valuable deeply human capabilities may become.
Creative thinking. Empathy. Collaboration. Resilience. Ethical judgement. Communication. Curiosity. The ability to listen. The ability to change your mind. The World Economic Forum’s research on the future of work continues to identify both technological skills and human capabilities such as creative thinking, resilience, flexibility and agility as important in a rapidly changing labour market.
Schools therefore do not need to become less human in order to prepare students for an AI-rich future. Instead they need to become more human. A meaningful conversation with a teacher matters. Collaborating with classmates matters. Making something with your hands matters. Debating an idea matters. Experiencing failure matters. Changing your mind matters. Belonging matters.
AI may be able to simulate many aspects of human interaction. That makes authentic human connection something we should fiercely protect.
6. Agency
Perhaps this is the most important item in the toolkit. We should not educate young people simply to become users of AI systems created by somebody else. They should understand enough to question them, influence them and imagine better ones.
Learning and teaching should position students not merely as consumers of AI but as responsible users and potential co-creators, emphasising a human-centred mindset, ethics, AI knowledge and AI system design.
That shift matters. The most powerful question we can give students this year may not be: “How can AI help you complete this task?”
It may be: “What do you want AI to help humans become capable of doing?”
That is a question about agency, values and the future we want to create.
The backpack test
So perhaps before students and teachers pack their metaphorical backpacks this year, we need a slightly different checklist. Bring your laptop. Bring your AI tools. But also bring:
Curiosity to ask better questions.
- Scepticism to challenge convincing answers.
- Judgement to know what to delegate and what to think through yourself.
- Transparency about how technology contributed to your work.
- Humanity to do the things machines cannot meaningfully replace.
- Agency, the confidence to use technology deliberately rather than simply allowing technology to shape how you learn.
AI will undoubtedly become more capable during this school year. The technology our students encounter in June 2027 may be substantially different from what they encounter in August 2026. That makes trying to prepare students for every new AI tool an impossible task. Instead, perhaps our responsibility is to equip them with something more enduring:
“The capacity to decide when to use AI, when to question it, when to put it aside and when to trust their own minds.“
That might be the most important toolkit we can give them.








