A practical guide for teachers who want real thinking, not robot homework
Whenever I walk into a teachers’ CBP, the first ten minutes are always the same. Someone sighs, someone else rolls their eyes, and then the chorus begins:
Students are submitting AI-made assignments… We have to evaluate this junk… Their basic reading, writing, and arithmetic are going downhill…
I get it. AI feels like that uninvited guest who shows up to the staffroom, finishes all the work, and still takes the credit. But here’s the twist: AI didn’t make students weaker. It only revealed which of our tasks weren’t asking them to think in the first place.
If a chatbot can score full marks on an assignment, that assignment wasn’t measuring thinking. It was measuring typing.
So instead of fighting AI like it’s Voldemort, here’s a smarter move: design tasks where AI can only support learning, not steal it. And yes, we’ll sprinkle a few puns along the way to keep things lively—because teachers deserve joy too.
Let’s break down 17 tasks AI can’t fake, dodge, or charm its way through.
1. Make AI the debate opponent
Students argue with a bot in real time. Great practice… and zero snack breaks required.
2. Bring in raw primary sources
Original letters, maps, court documents. AI can’t hallucinate its way through evidence it hasn’t seen.
3. Personal reflection prompts
When the question is about their life, their experience, and their learning, AI just awkwardly coughs in the background.
4. The counterclaim challenge
Students dismantle opposing arguments. Logical cardio at its best.
5. Identity-driven tasks
Culture, memory, family, belief—this is the territory AI can’t enter without getting lost.
6. Case studies with missing pieces
A real scenario, one key detail missing. Students must infer. AI hates incomplete puzzles.
7. Mid-task reasoning pauses
Stop them and ask why they made the last choice. No hiding behind perfect paragraphs.
8. Contradictory data sets
Give two reliable sources that disagree. Students must decide who earns their trust. (Spoiler: not AI.)
9. Student-designed experiments
They choose the question, method, tools, and interpretation. AI becomes the lab assistant, not the scientist.
10. Real-world interviews
No bot can replace the wisdom of a grandmother, carpenter, or local shopkeeper.
11. Reverse engineering assignments
Show the final answer. Ask them to unravel the steps behind it. A little Sherlock Holmes in every child.
12. Theory meets real life
Students connect classroom ideas to home, street, or community. Ground reality wins every time.
13. Build-your-own analogies
If they can compare democracy to a lunch queue, they understand it.
14. Ethical dilemmas
No right answer. Lots of thinking. Zero space for AI shortcuts.
15. Reflection checkpoints
Break long tasks into smaller intellectual pit stops. Students reveal their thinking, not their typing speed.
16. Local context projects
Neighbourhood issues, cultural practices, school challenges. AI wasn’t there—your students were.
17. Peer reviews with student-made rubrics
When students evaluate each other, judgment becomes the lesson.
Why these tasks matter
They build judgment, not just output.
They nurture curiosity, not compliance.
They demand insight, not copy-paste bravado.
And they turn AI from a threat into a tool.
Instead of asking “How do we stop students from using AI?”, we land on “How do we make them smarter with it?”
That’s the shift.
Not AI-proof learning.
AI-enhanced thinking.