Question quality critique

Have your own questions assessed and sharpened, so you stop asking prompts that only produce summaries and start asking ones that force thought.

When to use this: At the end of a session
Learning impact: Clear help, modest gain
Learning evidence: Credible with caveats
Where you are with the topic: Checking readiness
How often to use it: Worth repeating often

When to use it

Your chat history is mostly “explain X” and the answers feel like summaries.
Your study block is ending and you want sharper questions for tomorrow.
You caught yourself asking prompts that only dump information.

What to ask AI

Type something like: “Here are questions I wrote about [topic]: 1… Critique each on breadth, vagueness, and whether it forces thought.”

Ask for a rewritten version of each weak question, then improve them yourself further.

Keep choosing which questions you will actually use.

How it helps

Question quality shapes what kind of thinking you get back, including from yourself.

Dump-seeking questions produce summaries. Precise questions force decisions, comparisons, and evidence.

Critiquing your own questions is metacognition: you improve the instrument, not only the answers.

How to run it

Write five questions about a real topic you are studying.
Ask for critique: too broad, vague terms, not answerable, dump-seeking.
Rewrite the weak ones so each forces a decision or explanation.
Use the improved set in a real study block.

Example exchange

You

I drafted two study questions: “Explain photosynthesis,” and “Why must light reactions precede the Calvin cycle?” Which is better for real thinking?

Start with the first. What’s wrong with “Explain photosynthesis” as a prompt?

You

It’s a dump request, too huge, no decision. The second one is causal and bounded.

Good. Rewrite the first into a mechanism or decision question.

You

If a plant has CO₂ and water but no light, which products run out first, and why?

That forces sequencing and dependency. Use that one.

Copyable prompt

Here are questions I wrote about [TOPIC]:
1 …
Critique each on: (a) too broad / dump-seeking, (b) vague terms, (c) not
answerable, (d) whether it forces me to think. Suggest a sharper rewrite for
each weak one, then wait; I'll revise.

The Tell

If your chat history is still mostly “explain X” after a critique session, the critique did not change your habits.

If the AI rewrites everything and you paste blindly, you practised outsourcing, not question craft.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Improving question quality is metacognitive and links to learner-generated questions research: better questions support deeper processing.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI critique of questions can be generic. Judge rewrites against your real task. Delivery is speculative.