What will matter in education from 2027 to 2030? A useful starting point is to ask what students must still be able to do when their software changes. Explaining evidence, making a decision under uncertainty and applying a concept in a new situation are durable goals. The tools that help students practise those skills can change.
Teach AI literacy alongside subject knowledge
Ask students to compare an AI explanation with a textbook or primary source. Have them mark claims that can be checked, identify any missing assumptions and rewrite the answer in their own words. A geography class might compare an AI summary about a flood with an official map. A programming class might run tests against a generated solution and explain one case it missed.
Schools can set expectations by task. A tool may be appropriate for brainstorming or translation and unsuitable for a closed-book assessment. Explain the boundary before an assignment begins. UNESCO’s guidance on generative AI in education provides a starting point for human oversight, inclusion and privacy.
Assess the process as well as the result
A polished final submission tells a teacher little about how the student worked. Ask learners to bring a draft, describe one decision and revise something after feedback. Combine a portfolio piece with a short individual explanation or a new example solved in class. Use a simple rubric that rewards accuracy, evidence and the ability to correct mistakes.
Make flexible learning genuinely accessible
Combine a live discussion with a small downloadable practice set and written feedback. Offer captions or transcripts when possible, and make a low-bandwidth version available before assuming everyone can join video calls. If schools try tutoring or automated feedback, compare how learners do on the same independent task before and after the pilot. Check who could participate and who needed an alternative.
Plan a small experiment for the next school term
Choose one topic, one learning outcome and one measure of progress. For example, ask students to defend a claim with two reliable sources. Teach the skill with a worked example, give guided practice, and revisit it with a new topic two weeks later. Keep the parts that help students transfer the skill. For a way to adjust practice to each learner, see The Power of Personalized Learning.

