Most policies now ask you to declare it, not hide it.
Universities and journals have largely moved from banning AI assistance to requiring you to disclose it. This builds a clear, honest statement from what you actually did — no account, no upload, and nothing leaves your browser.
What did you use it for? (choose all that apply)
Wording
Your disclosure statement
I used ChatGPT to edit and proofread text written by the author. The author reviewed, edited and verified all resulting content and takes full responsibility for the accuracy and integrity of this work.
Reference-list entry
OpenAI. (2026). ChatGPT [Large language model].
This runs entirely in your browser. Nothing you type is sent to us or stored anywhere.
Will your writing pass an AI detector?
Disclosing AI use is good practice — but teachers and editors also run detectors. Scan your text with Plagly’s full AI + plagiarism report before you submit.
See Plagly plansWhy disclose at all
It is usually the actual rule
Most institutions that have updated their policies now permit AI assistance and require it to be declared. Undisclosed use is what gets treated as misconduct — not use itself. Read your own department's wording, because the details vary.
It protects you from a false positive
Detectors flag human writing too, especially short, formal, heavily-edited work by non-native speakers. A disclosure you wrote in advance, alongside your drafts, is far better evidence than an argument made after an accusation.
It is a low-cost habit
Two sentences at the end of a document. The cost of including it when it was not required is nothing; the cost of omitting it when it was required can be a misconduct case.
What this tool is and is not
What it does
- +Assembles a clear statement from your own answers, deterministically. It does not call a model, so it cannot invent a claim about your process.
- +Offers wording that fits a plain statement, an APA- or MLA-style acknowledgement, or a methods section.
- +Adds a reference-list entry for tools whose publisher we can identify, and stays silent for ones we cannot rather than guessing.
What it is not
- –Not legal or academic advice, and not a substitute for your institution's own template. Where they publish required wording, use theirs.
- –Not a way to make undisclosed use acceptable retroactively. A disclosure describes what you did; it does not change it.
- –Not a guarantee of acceptance. Some journals and courses still prohibit generative AI for certain tasks, and a disclosure does not override that.
Frequently asked questions
Do I need to disclose using AI just for spellcheck or grammar?
Usually not. Most policies distinguish between longstanding assistive tools — spellcheck, grammar suggestions, reference managers — and generative tools that produce content. If you asked a model to write, rewrite, translate or summarise, that is the kind of use policies mean. If you are unsure, disclosing costs you nothing.
Where does the statement go?
Commonly at the end of the document before the references, in an acknowledgements section, or in the methods section for research papers. Some institutions ask for it on a cover sheet or in a submission form field instead. If yours specifies a location, follow that.
Should I list the AI as a co-author?
No. The major publishers and style guides are consistent on this: an AI system cannot be an author because it cannot take responsibility for the work or approve the final version. Disclose it as a tool you used, which is what the methods-section wording here does explicitly.
Is anything I type here uploaded?
No. The statement is built in your browser with plain JavaScript. There is no request to our servers, nothing is logged, and you can disconnect from the internet and it will still work.
My institution has its own required wording. Should I use this?
Use theirs. Where an institution publishes a required form of words, that is the one that counts. This tool is for the common case where you are told to disclose but not told exactly how to phrase it.
Checking the document itself?
Our detector shows sentence-level results, not just a percentage.
Try the AI detector