Teaching AI literacy: 10 classroom activities for primary and secondary
Ten practical AI literacy lessons, from unplugged primary activities to secondary tasks on evaluating chatbot answers, bias and responsible AI use.
Key takeaways
- AI literacy can be taught in short, repeatable activities within existing subjects.
- Unplugged activities help younger students understand data, patterns and training without needing accounts.
- Secondary activities should focus on evaluating AI output, verifying claims and deciding when to use AI.
- Each activity is mapped to one of the four AILit domains: Engage, Create, Manage and Shape AI.
- Check the age requirements and data terms of any AI tool, and demonstrate from the teacher’s account for younger students.
These activities are designed to fit into ordinary lessons. Each one is mapped to a domain from the OECD–European Commission AI literacy framework. Timings are approximate; adapt the content to your curriculum and learners.
Primary activities
1. Spot the AI (Engage · 20 minutes)
Show pictures of everyday technology: a voice assistant, a calculator, a video recommendation screen, a light switch, a photo filter, a map app. In groups, students sort them into “uses AI”, “does not use AI” and “not sure”, then explain their reasoning. Discuss the clue: AI systems learn patterns from data and make predictions, while a calculator follows fixed rules.
2. Teach the machine (Shape · 30 minutes)
Give pairs a set of animal picture cards. One student is the “machine” and learns only from the examples the “trainer” shows: “These are birds, these are not.” Then test the machine with new cards, including a penguin and a bat. Discuss what happened when the training examples were limited. This is how data shapes what AI can and cannot recognise.
3. Human or machine? (Engage · 20 minutes)
Read two short descriptions of the same picture, one written by the teacher and one produced by an AI tool before the lesson. Students guess which is which and explain their clues. Talk about what AI does well and what it misses, such as feelings, context or what matters most.
Middle and secondary activities
4. The confident wrong answer (Engage · 25 minutes)
Before class, ask a chatbot a question from your current topic and collect answers that contain a subtle error. Students compare the answer with their textbook or notes, find the mistake and explain why a fluent answer can still be wrong. End by writing a class rule for checking AI answers.
5. Check, compare, cite (Engage · 30 minutes)
Give students three claims generated by AI on a topic you are studying. For each, they find a trustworthy source that confirms or contradicts it, using lateral reading: leave the page and see what other sources say. They record the source and a verdict: supported, partly supported or not supported.
6. Summary versus source (Engage · 30 minutes)
Students read a text closely, then compare it with an AI summary. What did the summary leave out, simplify or get wrong? Which details would matter for an exam answer? This builds close reading as well as AI literacy, a direct response to the PISA 2025 reading findings.
7. Draft first, then AI (Create · 40 minutes)
Students write a first draft on their own. They then ask AI for three suggestions to improve it, and decide which suggestions to accept, change or reject. The final piece includes a short rationale explaining each decision. The thinking stays with the student, and the reflection shows it.
8. Prompt remix (Create · 25 minutes)
Start with a simple prompt, such as “Explain photosynthesis.” Groups change one element at a time: the audience, the format, the length or the example. Compare the outputs and discuss how context changes the result, and why the output still needs checking.
9. Who should do this task? (Manage · 20 minutes)
Give students cards describing tasks: brainstorming project ideas, writing a personal reflection, checking spelling, solving a homework problem, translating a message to a grandparent, choosing a career. They sort each into “me”, “AI can help”, “AI and me together” or “not AI”, and justify their choices. There is no single right answer; the discussion is the learning.
10. Our class AI charter (Shape and Manage · 40 minutes)
Using what students have learned, co-create a class agreement: when AI is welcome, when it is not, how to show where AI was used and how to check its output. Revisit the charter each term. Connect it to your school’s AI policy.
Assessing AI literacy
- Ask students to explain their reasoning, not only their conclusion.
- Collect short reflections: What did AI get right? What did you change and why?
- Look for transfer: do students check AI output in other subjects without being prompted?
Frequently asked questions
How do you teach AI literacy to primary students?
Use unplugged activities such as sorting everyday technology or acting out how a machine learns from examples. Teachers can demonstrate AI tools from their own account rather than asking young students to create accounts.
What age should students start using AI tools?
Check each tool’s terms. UNESCO’s 2023 guidance on generative AI in education suggests a minimum age of 13 for classroom use of these tools.
How long should an AI literacy lesson be?
Short activities of 15 to 40 minutes, repeated across subjects, often work better than a single long unit.