Pedagogy guide

Why this is not a first-draft generator.

The toolkit is designed to protect the student’s first struggle with ideas, structure and wording because that struggle is part of learning.

Pedagogical reason

The first struggle matters.

The toolkit is mainly designed for diagnosis, revision, explanation and practice after the student has made an attempt. If AI does too much at the start, it can remove the very struggle that builds understanding, judgement and ownership.

Why the first struggle matters

Writing is not just something you do after thinking. It is often how the thinking happens. When you try to name your point, order your evidence, explain a link or choose a word, you are working out what you understand.

The first attempt is where you notice what is clear, what is vague, and what you cannot yet explain. If AI makes the early choices for you — the argument, the structure, the examples or the wording — you may end up with a smoother draft but less practice in the thinking the assignment is meant to develop.

That does not mean you can never use AI early. It means the safest early use is the kind that helps you think: asking questions, unpacking a brief, practising on made-up examples, or testing an idea you have already started to form.

Why not use AI to create the first draft?

Initial drafting is not just a production task. It is where students decide what they think, notice what they do not understand, choose a structure, test language, and discover gaps in their reasoning. Those difficulties are frustrating, but they are also part of learning.

If an AI tool supplies the first structure, argument or wording too quickly, the student may end up editing something they did not really think through. That can make the work smoother while weakening the learning.

Scaffolding can become substitution

Scaffolding is useful when it helps a learner do something they could not yet do alone. But if the scaffold makes the key decisions for them, it stops being support and becomes substitution.

This is why the toolkit usually asks for the student's own text, plan, feedback or attempt before giving detailed help.

What early-stage use is acceptable?

  • Asking what an assignment brief is asking you to do.
  • Practising with a made-up example before applying the idea to your own work.
  • Using Socratic questions to explore what you already think.
  • Checking whether a research question is too broad or too vague.
  • Planning how to approach a task without asking AI to write the answer.

Better default

Try first. Then use the toolkit to diagnose, learn, revise and reflect.

This keeps the difficult thinking with the student while still making AI useful as a tutor.