Prompt literacy vs question literacy: the skill that actually matters

Prompt engineering has become the buzzword everyone reaches for, but it teaches a narrow skill — how to phrase a request — while the deeper and much harder skill is knowing what's actually worth asking in the first place. That second skill, question literacy, is prior to prompt literacy in every meaningful sense, and mistaking one for the other is a real and common error in how AI use is being taught right now.

What prompt literacy actually is

Prompt literacy is real. It covers things like being specific rather than vague, giving relevant context up front, specifying the format or length you want, and iterating when the first response misses the mark. These are genuinely useful moves, and a student who learns them will get noticeably better outputs from any AI tool than one who doesn't.

But look closely at what's being taught: it's entirely about how to phrase something you already know you want to ask. It assumes the hard part — figuring out what to ask — is already done. That assumption is where the whole framing goes wrong for education specifically, because for a KS3 student the hard part usually isn't phrasing. It's knowing what's actually missing from their own understanding.

What question literacy actually is

Question literacy is the skill of accurately sensing your own gap — knowing not just that you're stuck, but where, specifically, and what kind of question would actually close that gap. It's the difference between a student who can only say "I don't get it" and one who can say "I understand how to factorise when there's no constant term, but I lose the plot as soon as there's a negative coefficient."

That second version is a genuinely useful question. It's specific enough to act on, and — crucially — writing it required the student to already understand their own confusion well enough to locate it precisely. No amount of prompt-phrasing skill gets you there, because the phrasing skill starts after you already know what to ask.

This is exactly what the Ask Ladder is a tool for: noticing which level of question you're actually capable of asking on a given topic right now. A student stuck at "what does this word mean" (Locating) hasn't yet built the self-awareness to ask "where exactly did my reasoning go wrong" (Diagnostic) — and that progression has nothing to do with prompting technique. It's built through repeated practice at noticing your own confusion, which is a slower and less flashy thing to teach than prompt syntax, but it's the thing that actually matters.

At a glance

Why the smaller problem gets all the attention

Prompt literacy gets disproportionate attention for an understandable reason: it's easy to teach, easy to demonstrate, and easy to feel like progress on. A one-hour workshop on "how to prompt ChatGPT" produces visibly better outputs by the end of the hour. That's satisfying, and it's not nothing.

But it solves the smaller problem. A student who's brilliant at phrasing requests and has no sense of their own gaps will still ask the wrong things — fluently. They'll get excellent answers to questions that don't actually address what they don't understand, and walk away no closer to real comprehension, just more confident that they are. This is a specific and dangerous version of the Answer Trap: a well-phrased question that skips the harder work of figuring out what needed asking.

Schools and parents who focus a curriculum or a family's AI conversation entirely on "how to prompt well" are, without meaning to, teaching the easier half of the skill and leaving the harder half untouched.

Building question literacy in practice

Question literacy doesn't come from a lesson on syntax — it comes from repeated, low-stakes practice at noticing your own confusion and naming it. A few practical routes into that:

• Practise the Ask Ladder explicitly, on ordinary homework, not as a special exercise — after finishing a piece of work, ask "what level of question did I actually ask about this today?"
• Use the First-Guess Rule before consulting anything — committing to a guess forces you to locate what you actually think you know, which is the raw material question literacy is built from.
• Keep a running note of questions you swallowed rather than asked — see Question Debt — because reviewing what you didn't ask is often more revealing than reviewing what you did.
• Let a tutor or parent model it, the way aitutors.me's subject tutors do — Professor Pi's hint ladder and Professor Newton's Predict-Observe-Explain both work by making the student locate their own gap before offering anything, rather than answering a vague "I don't understand" outright.

FAQ

What's the difference between prompt literacy and question literacy? Prompt literacy is the narrow, teachable skill of phrasing a request to an AI tool to get a useful output — it's about HOW to ask. Question literacy is broader and comes first: knowing WHAT is actually worth asking, which requires an accurate sense of your own gap in understanding.

Can prompt engineering be taught in a single lesson? Largely, yes — the core moves (be specific, give context, ask for the output format you want, iterate on the response) can be taught and practised in an afternoon. That's exactly why it's a smaller problem than it's often treated as: it's easy to teach because it's narrow.

Why can't you substitute prompt skill for question literacy? Because a perfectly phrased prompt for the wrong question still produces the wrong outcome. No amount of skill in HOW to ask fixes not knowing WHAT you actually need to know — that only comes from an accurate sense of your own gap, built by practising the Ask Ladder over time.