Teach AI: building pupil AI literacy in UK schools

Why teaching AI is now a pedagogical decision

Pupils are already using AI. Pretending otherwise is the riskiest option. Teaching AI well means giving pupils the conceptual and critical scaffolding to use it as a thinking tool rather than a shortcut around thinking.

What "AI literacy" actually covers

• How generative AI works at a high level — pattern, not understanding
• Where AI fails: hallucination, bias, plausible-sounding wrongness
• Academic integrity and disclosure expectations
• Data privacy — what not to type into any AI tool
• Metacognition: "how do I know this answer is right?"

How to teach AI without de-skilling pupils

1. Always require pupils to do the thinking first, then check with AI
2. Have pupils critique AI output as a regular routine
3. Use AI to expose misconceptions, not to deliver answers
4. Make disclosure ("I used AI to…") a normal part of submissions

Primary, secondary and post-16: where the emphasis sits

In primary, the message is that AI can be wrong and is not a person. In secondary, the emphasis moves to interrogation, integrity and disclosure. Post-16, pupils need explicit practice in using AI as a research and reasoning partner without offloading the reasoning.