← All guides
Writing7 min read

How to Prove You Didn't Use AI on Your Essay

Flagged by an AI detector for work you wrote? What to do in the first hour, where to find your version history, and how to prepare for the meeting.

A stack of dated drafts beside a version history timeline, each point a saved stage of the same document growing over time.

You wrote it yourself. An AI detector says otherwise, and now there is an email asking you to come in and talk about it. The panic is understandable. It also helps to know how these cases actually work: they turn on evidence and on conversation, and a student who did the work usually holds more evidence than they realise.

You are far from the only one worried

In the Higher Education Policy Institute's 2026 survey of UK undergraduates, 42 percent said fear of being accused of cheating made them less likely to use AI at all. A 2026 survey reported by Inside Higher Ed found nearly three quarters of US college students at least moderately worried about a wrongful accusation. The worry is reasonable. A detector reports a probability, not a finding, and careful human writing can read as machine-like to it.

Read nextWhy AI Detectors Flag Writing You Actually WroteAI detectors flag real student writing every day, and it is worst for non-native speakers. Here is why it happens and what to do if it happens to you.
→

What reviewers actually weigh

A score should start a conversation, not end one. England's Office of the Independent Adjudicator, which reviews student complaints against universities, has published a casework note on AI misconduct. Its position is that the university carries the burden of proof, and the evidence that helps students is the ordinary evidence of doing the work: notes, drafts, file history the student can explain, and being able to discuss the content in a meeting.

Courts look at process too. In January 2026 a New York court overturned Adelphi University's misconduct finding against a student that had rested on a detector report, because the school ignored the evidence the student offered and did not follow its own rules. The record was cleared, but it took a lawsuit to get there. The practical goal is to settle things at the first meeting, with your evidence already in hand.

The first hour

A reply you can adapt

Dear Professor [name], thank you for letting me know about the concern with my essay. I wrote it myself, and I would welcome the chance to show you how. I have my notes, earlier drafts and the document's version history, and I am happy to talk you through my argument and my sources. Could you tell me what the concern is based on, including any detector report, and when you would be free to meet? Kind regards, [your name]

Ask for a copy of the misconduct procedure your school follows, too. It tells you what happens next, who decides, and whether you can bring an adviser.

Evidence you probably already have

Most of the proof of honest work is created automatically while you write. You only need to know where to find it.

If you use assistive technology such as dictation or spelling support, or you have a learning accommodation, say so early. It can explain patterns in your writing, and the accommodation may apply to the meeting itself.

Be ready to talk it through

The strongest evidence is often you. A common next step is a short meeting about the work, sometimes called a viva. The University of Bath's guidance puts the purpose well: the conversation should test whether a student understands how they reached their conclusions, not whether they can recite the text.

If you did use AI for part of it

Say so, plainly and early. Many courses allow some AI help, such as brainstorming or checking grammar, as long as you disclose it. A clear statement of what you used and what you wrote yourself is far easier to discuss than a denial that later comes apart. If your course has an AI policy, match your statement to its wording.

Read nextHow to Disclose AI Use in an AssignmentMore courses now allow AI and require you to say how you used it. What a disclosure statement is, what to put in it, a template you can adapt, and the two mistakes that cause trouble.
→

One thing never to do

Do not manufacture a process after the fact. There are tools that retype finished text into a document to create a fake revision history. Using one turns a misunderstanding into real misconduct.

Keeping a record next time

The easiest defence is the one you build as you go. Write in one document rather than pasting in from elsewhere, keep your notes alongside it, and save a copy at each stage.

Our editor does this for you. It keeps a private record of how your document grew: time spent writing, what was typed and what was pasted, and a timeline of the draft taking shape. You can share it as a link that lets a reader check the document they have matches the record, and the Defend tab turns your draft into the questions a marker is likely to ask, so you can practise the answers.

We are careful about what that record claims. It shows how a document was written in our editor. It is not a certificate that no AI was used anywhere, and no record can honestly be one. What it gives you is organised, dated evidence, which is what these conversations turn on.

Read nextWhat the Turnitin AI Report Actually Shows Your ProfessorThe AI writing indicator is not the similarity score. What your professor sees, what the percentage means, and what Turnitin says about its own limits.
→
Read nextWhy AI Detectors Flag Non-Native English Writers (and How to Protect Your Work)AI detectors flag writing by non-native English speakers far more often. A Stanford study found the majority of their essays wrongly flagged. Here is why, and what to do.
→

Paste your essay and practise the questions a marker is likely to ask, drawn from your own lines.

Prepare to explain your work →

More guides