Why disclosure matters#
Disclosure lets the person assessing your work judge it fairly. A marker who knows you used an AI tool to tidy grammar reads your essay differently from one who has to guess. In research, readers need to know how a result was produced in order to trust and reproduce it.
It also protects you. If questions are raised about your work, a clear, accurate statement made at the time, backed by notes and drafts, is far stronger evidence than anything you can assemble afterwards. Trying to hide AI use, or to make AI-written text look human-written, turns a question about tool use into a question about honesty.
First, find the policy that applies#
Rules on AI vary widely, between institutions, between departments and even between assignments on the same course. Some allow AI for any purpose with disclosure, some allow it only for specific tasks, and some do not allow it at all. Before you write anything, look for:
- the assignment brief or module handbook,
- your department's or programme's guidance on AI,
- your institution's academic integrity policy,
- for publications, the journal's and publisher's author guidelines.
The most specific rule wins. If the brief says "no AI tools", a general university policy that allows them does not override it. If you cannot find a rule, or it is unclear whether your use is covered, ask your instructor or supervisor in writing before you rely on the tool. Their answer is worth keeping.
What a good disclosure includes#
Most institutional and publisher guidance asks for some version of the same elements. If your policy specifies a format, use it. If not, cover these:
| Element | What to say | Example |
|---|---|---|
| Tool | The name of the tool, and the version or date if you know it | "MoreMind" or "ChatGPT (GPT-4o, accessed March 2026)" |
| Purpose | The task you used it for | "to check reference formatting" |
| Extent | Which parts of the work were affected, and how much | "in the reference list only" |
| Your role | What you did with the output | "I reviewed and corrected every entry" |
| Responsibility | That you are accountable for the final content | "I take full responsibility for the content of this work" |
Be specific. "I used AI to help with this essay" tells the reader almost nothing. "I used [tool] to suggest corrections to grammar and spelling in my final draft; I accepted some suggestions and rejected others, and did not use it to generate content" tells them what they need to know.
Some courses also ask you to attach prompts and outputs, or to cite AI-generated text that you quote. APA and MLA have each published guidance on citing generative AI; if you quote or paraphrase AI output, check the current guidance for your style. Citing and disclosing are different things: a citation credits specific material, a disclosure describes your process.
Where to put the statement#
- Coursework: where the brief says. If it says nothing, a short paragraph headed "Use of AI tools" at the end of the work, before the reference list, or on a cover sheet, is clear and easy to find.
- Theses and dissertations: often in the acknowledgements or a declaration page, and sometimes in the methods chapter if AI was part of how data were processed. Check your graduate school's regulations.
- Journal articles: usually the methods section (when AI was part of the research process) or the acknowledgements (for writing or editing help), following the journal's instructions.
Example statements by use level#
These are starting points. Edit them so they describe exactly what you did, no more and no less, and replace them with your institution's wording if it provides one. A longer set, with notes on adapting each, is in the printable AI-use disclosure templates.
If AI played a bigger part#
The more an AI tool contributed, the more your statement needs to say. If you used one to help analyse data, transcribe interviews, translate sources or generate code, describe the task, the tool and how you checked the output, in the same way you would describe any other method or software.
If you are not sure whether a use crosses a line your course draws, disclose it and ask. Describing a use your instructor considers minor costs you nothing; leaving out one they consider significant can cost a great deal.
Keep records that back it up#
A disclosure is more convincing when you can show your process. Keep:
- dated drafts or version history (most word processors and cloud drives keep this automatically),
- your notes and source matrix,
- prompts and outputs, if you used a chat-based tool and your policy asks for them,
- any written permission or clarification from your instructor.
You will rarely need them. If you do, they settle questions quickly.
Mistakes to avoid#
- Under-describing. Saying "grammar checking" when the tool rewrote whole paragraphs is not an accurate disclosure.
- Over-claiming. Do not say you checked every source if you did not. Go back and check them.
- Copying a template unchanged. A statement that does not match what you did is worse than none.
- Relying on detectors. AI-detection tools are not a reliable measure of authorship, in either direction. An honest statement and a clear record of your process carry more weight than any detector score.
- Assuming a past rule still applies. Policies on AI are changing quickly. Check them for each new course or submission.
If you are unsure whether a particular use is acceptable, the scenario checker on the academic integrity page walks through common cases, and our use cases show where MoreMind's help stops and your work begins.
KEEP GOING
https://moremind.ai/en/guides/how-to-disclose-ai-use