QUESTION · FOR TEACHERS

How do you teach students to write good AI prompts?

Teach prompting as a learnable skill, not a trick. Have students keep a prompt log: the first attempt, what was wrong with the answer, and a sharper second attempt. Give them a simple checklist (goal, context, audience, format, constraints) and teach them to treat every first answer as a draft to question and improve.

Written by David Fineberg · Last updated

Why teach prompting at all?

A vague ask gets a vague answer, and students may conclude the tool is either magic or useless. Teaching students to prompt well builds the habit of deciding what they actually want before they ask. In the WISE framework this is the Instruct It pillar: the better the ask, the better the answer.

Prompting is only one part of AI literacy. For the wider picture, see AI literacy vs. prompt engineering.

Start with a prompt log

The most useful classroom tool is a prompt log, a simple table students fill in as they work. It makes thinking visible and gives the teacher something real to assess. Each row has four parts:

  1. First attempt: exactly what the student typed.
  2. What was wrong: what the answer missed, got wrong or made too general.
  3. Second attempt: a sharper prompt that fixes those problems.
  4. What changed: a sentence on why the second answer was better or worse.

Comparing two real attempts gives students something concrete to learn from, which a handed-out formula does not. Assessing the log also means assessing the thinking.

What makes a good prompt?

Give students a short checklist they can keep beside the screen. A good prompt usually includes:

  • Goal: what the student is trying to do or understand.
  • Context: the grade level, the topic, what they already know.
  • Audience: who the result is for.
  • Format: a list, a short paragraph, a table, a set of questions.
  • Constraints: length, reading level, what to leave out.
  • Reasoning or questions: ask the tool to explain its thinking, or to ask the student questions first.

Students need not use all six every time; they just notice which are missing when an answer disappoints.

A before and after example

Here is a prompt pair on a Grade 8 science topic.

Before: "Tell me about the water cycle."

After: "I am a Grade 8 student studying the water cycle for a quiz. Explain evaporation, condensation and precipitation in plain language, using a lake near a mountain as the example. Use a short numbered list, no more than 150 words, then ask me three questions to check what I understood."

The second prompt names the goal, the context, the format and the length, and it asks the tool to quiz the student.

Teach students to let the tool ask the questions

Many students do not know what to include in a prompt. Show them a stronger move: ask the tool to interview them. For example: "Before you answer, ask me up to five questions about what I need." This turns a one-shot request into a short conversation and often surfaces what the student had not thought through.

Also teach that the first answer is a draft. Students can ask for a simpler version, a counterargument, or the weakest part of the answer. Then they check the claims against real sources, as described in why AI makes things up.

Prompting does not replace knowing the subject

A student who knows little about a topic cannot tell whether an answer is good. Strong prompts often come from students who understand the material well enough to notice what is missing. Make this explicit: prompting skill sits on top of subject knowledge and does not replace it. Where the goal of a task is to build a skill such as writing an argument, decide in advance what the student does alone and what the tool may help with. That is the Wield It pillar, and the rules are covered in Can students use AI for schoolwork?

Where to start

  1. Pick one ordinary assignment and decide what AI may and may not do.
  2. Give students the checklist and a blank prompt log.
  3. Ask for two attempts and a short reflection on the difference.
  4. Discuss a few logs as a class, without naming students.

The Prompt-Portfolio Set includes a redesigned assignment, a four-part portfolio and a sample rubric. Learn more under free resources. The WISE framework explains how the four pillars fit together.

Quick answers

Should students memorize prompt formulas?

No. Formulas help at first, but the lasting skill is noticing what is missing from an answer and adjusting the ask. A prompt log builds that habit better than memorizing a template.

Does a better prompt guarantee a correct answer?

No. Even a well-written prompt can produce confident but false information, so every factual claim still needs to be checked against a real source.

Free classroom-ready resources

Handouts, a sample school AI policy, a prompt-portfolio set and a printable WISE Compass poster. Enter your email to get them, plus new resources as they are added.

Get the free resources
Free · unsubscribe anytime
ABOUT THE AUTHOR

David Fineberg is an educator and AI-in-education consultant based in Montreal, working with English and French school communities. He is the author of Teaching in the Age of AI: The WISE Canadian Classroom. He offers staff workshops, consultations and talks through Edify AI.