Teaching students to prompt is teaching them to think.
A good prompt is a well-formed question. Treat prompting as a writing skill, not a trick, and the AI lesson turns into a thinking lesson.
SKShreyas KoushikAll Things Classroom
There is a version of "teaching prompting" that is just teaching shortcuts: paste this, add that magic phrase, get a better answer. It is not useless, but it is shallow, and it ages fast because the tools change. The version worth doing is different. It treats a prompt as what it actually is — a question with its constraints made explicit — and it turns out that writing one well is most of the way to thinking clearly about a problem.
Four things a good prompt makes a student decide
- What exactly am I asking? Vague prompts get vague answers, which is a lesson about vague questions.
- What does the tool need to know that it can't guess? Context, audience, purpose, format.
- What would a good answer look like? If a student can't say, they can't judge the output either.
- What am I going to do with the result? Use it, check it, or push back on it.
None of these are AI skills. They are the same moves a student makes when planning an essay or designing an experiment. That is the point. The tool gives the thinking somewhere to go, and gives fast, visible feedback on whether the thinking was clear.
The prompt log
In our AI Literacy programme students keep a prompt log: the prompt, what came back, what they changed, and why. It is the most useful piece of evidence a teacher can collect, because it shows the reasoning, not just the result. A log where the second prompt is more precise than the first is a student learning. A log where every prompt is a single line is a student who needs a different task.
The best prompts read like briefs written by someone who already knows what good looks like.
Two classroom habits help. First, make students write their prompt before they open the tool, on paper if you like. Second, ask them to predict the answer. The gap between prediction and output is where the learning is — sometimes the model was wrong, sometimes the student was, and the class gets to work out which.
Finally, prompting is a place to teach restraint. The fluent prompter who knows when to close the laptop is the goal; the fluent prompter who can't is a problem we've created. A prompt log with the entry "didn't use it — this was mine to do" is a good day.
Further reading
Written by
Shreyas Koushik
All Things Classroom
Writes about learning design, AI in the classroom and what actually works with students.
