GRASP

Should I be using GenAI to help write my thesis?


September 2nd, 2026, by Karen Miller Tag(s): GenAI, Academic Writing

GenAI can help at almost every stage of writing a thesis - generating ideas, structuring your argument, tightening prose, even reflecting on your own progress. Used well, it can save time and reduce the friction of getting started on writing tasks.

However, writing a thesis is not just about producing a document - it’s how you become a researcher. Some of the effort GenAI can potentially take off your hands is effort you actually need to make yourself in order to develop the skills you are expected to have as a researcher.

This post will help you think about ways to use GenAI in your writing which can develop your critical thinking skills rather than undermine them, and ensure you use GenAI appropriately and effectively. It aligns with Curtin’s recently updated Guidelines for the use of AI in Research.

What’s at stake

Through the struggle of writing you build three things: your skills, your understanding of the topic, and your accountability for the work.

Your skills. Writing, planning, and revising are how you practise the abilities you’ll need as an HDR graduate - constructing an argument, evaluating evidence, working through a problem before the answer is obvious. If GenAI does that work for you, you’re left with a skills gap that shows up the next time you have to do it alone.

Your understanding of the topic. For most people, thinking happens through writing, not just by recording thoughts already formed. You often don’t know what you think until you’ve tried to explain it clearly and revised the explanation when it didn’t hold up. If GenAI generates the explanation instead, you can end up with a fluent paragraph that sounds like your position without you ever having gone through the process of forming it. That matters not only for your own critical thinking ability, but also when a supervisor or examiner asks you to defend your position.

Your accountability. As a researcher you’re answerable for the claims and interpretations in your thesis, not just the wording. The issue with presenting GenAI-generated ideas as your own thinking isn’t plagiarism in the traditional sense as AI has no authorship to take credit from. It’s a question of whether or not the thesis accurately represents your own thinking.

Using GenAI as a ‘thinking partner’

This doesn’t mean you need to avoid GenAI altogether in your writing. An effective approach is to treat GenAI as a thinking partner rather than a thinking substitute. A thinking partner helps you reason - it doesn’t reason in your place.

In practice, that means having a back-and-forth conversation with GenAI rather than entering one polished prompt and receiving an answer that you accept. You need to:

  • Ask it to justify a suggestion.
  • Push back if something doesn’t sit right.
  • Ask it to explain a term rather than assuming you’ve understood.
  • Say what you think before asking what it thinks.

The hidden risks

When you and GenAI are reasoning together, the line between your input and its input can blur without you noticing - and a fluent, productive conversation is not a reliable sign that your own thinking is keeping pace. Watch for:

  • Assuming the AI shares your conceptual frame when it doesn’t.
  • Letting a term slide unquestioned because stopping to ask would break the flow.
  • Being persuaded by how a response reads, rather than by your own reasoning confirming it.
  • Realising, in hindsight, that your framing of the problem shifted during the conversation without you deciding to shift it.
  • Noticing that something you’d flagged as “needs checking” became “settled” because you kept building on it.

Uses and risks across the writing process

GenAI can be useful at every stage of writing a thesis, but what “useful” looks like changes from stage to stage, and so does what can go wrong. It can also play quite different roles depending on what you’re asking of it, whether it is handling a bounded task you’ve set (proofreading, summarising, structuring), or contributing more directly to the thinking itself.

Stage High-value use Risk zone
Idea generation Question prompting, widening perspectives Letting GenAI choose the topic/angle for you
Planning Testing structure, comparing outlines Outsourcing the argument’s logic
Drafting Clarifying a sentence, checking cohesion Generating substantive content you haven’t thought through
Revising Comparing versions, catching clunky phrasing Polished prose masking a conceptual gap
Reflecting Asking it to find weaknesses in a finished piece Accepting feedback as authoritative without scrutiny

Strategies for putting effective GenAI use into practice

1. Build a prompt that helps catch the hidden risks.

Real-time prompting - asking GenAI to flag specific failure patterns as they happen - is a more reliable safeguard than trying to catch them yourself in the moment, when you’re least likely to notice. A good prompt names specific patterns rather than asking it to just “flag problems,” frames flags as questions rather than verdicts, and includes a check against you accepting a flag uncritically. Here’s an example you could paste into a chat or set as a standing instruction:

As we work together, watch for the following and flag them as a question, not a verdict, since I may not notice them myself in the moment:

  • If I accept a suggestion from you without saying what I actually think first, ask me what I think before we move on.
  • If I want to move straight to a final answer without hearing a counter-argument, ask whether I want to consider one first.
  • If you introduce a term or idea I might not know, check rather than assuming my silence means I’ve understood.
  • If a claim you make sounds confident but isn’t something I’ve checked myself, say so, and suggest I verify it.
  • If I seem to be mostly accepting and lightly editing your output rather than writing my own material, ask me to state, without looking at the draft, what I actually think the argument is.
  • If I accept one of these flags without actually testing whether it’s true, notice that too, and ask me to slow down rather than just deferring to the flag itself.

Adapt this, or add flags for your own known tendencies.

2. Write your own GenAI use statement.

Decide in advance, rather than in the moment, how you intend to use GenAI when writing your thesis as a personal position you could defend if asked. Cover three things: how you intend to use it (which tasks, which stages, why), where you draw the line, and how you’ll make sure the final thesis represents your own thinking rather than your editing of someone else’s. Revisit it as your project develops.

3. Document as you go.

It’s a good idea to keep a record of key prompts and outputs especially anything that shaped your research design, analysis, or interpretation as you are required to disclose your use in the thesis itself. This is much easier if you keep a running log as you work, rather than reconstructing it from memory at the end. An example disclosure statement is provided in the Guidelines and goes something like this:

“This thesis has benefited from the use of Generative AI tools (specifically [tool]) to improve the clarity, structure, and readability of certain sections. All substantive intellectual contributions, including the conception of ideas, research design, data analysis, and interpretation of findings, are entirely the author’s own.”

If your use goes beyond editing, say so explicitly.

4. Verify, don’t just accept.

GenAI output isn’t a neutral second opinion — it reflects the standpoint of its training data, including assumptions about what counts as good evidence or a sound argument. Critically evaluating it means checking those assumptions, not just checking facts.

5. Take extra care if English isn’t your first language.

Using GenAI for proofreading and grammar is acceptable and legitimate language support. Ensure you check a rewritten sentence against what you actually meant to say, not only against whether it reads well — this matters for everyone, but can be harder to catch if you’re still developing your English writing skills.

Final thoughts

It’s not necessary to completely avoid using GenAI for your academic writing, but it’s important to remain the author of your own thesis. The line between GenAI as a useful thinking partner and a substitute for your own thinking isn’t always obvious in the moment, and there are some strategies you can adopt to avoid the hidden risks. Creating a suitable risk detection prompt you can use at the beginning of any conversation with GenAI is an effective way to ensure GenAI can help you develop your thinking and writing skills, rather than undermine them.

Sources used

This guide draws on Shana Pote’s course on Governing Thought in the Age of AI: Diagnosing and  Countering Epistemic Failure Modes in AI-Engaged Scholarship as well as content from Curtin Library’s upcoming GenAI and Academic Writing workshop and Curtin’s Guidelines for the use of AI in Research.


Please make any anonymous comments/ feedback, or suggestions for further posts at this link. If you would like to get in touch, or write a post for the Ideas Hub blog, please email karen.miller@curtin.edu.au.


Image by Lone Thomasky & Bits&Bäume at Better images of AI shared under Creative Commons license 4.0