Writing Guides

Falsely flagged as AI-written? What to do next

Save your version history before anything else, find out exactly what is alleged, and explain your process in writing.

By aitextdetectly team2 min read

If a teacher, editor, or employer says a detector flagged your writing as AI, you aren’t the first, and the tool may well be wrong. Some writers get flagged more often than others: a 2023 Stanford study found that widely used detectors consistently misclassified essays by non-native English speakers as AI-generated (Liang et al., Patterns). That doesn’t settle your case. It’s worth knowing before the conversation starts.

Save your history before you change anything

Don’t edit the document yet. If you wrote it in Google Docs, open File > Version history > See version history. It lists timestamped versions and who made each change. Word keeps a similar list under File > Info > Version History for files saved to OneDrive or SharePoint. A document that grew over several sessions, with deletions and rewrites along the way, looks very different from one that arrived in a single paste.

Collect what exists outside the file too: outlines, notes, sources you looked up, earlier drafts you sent to yourself or someone else.

Take a screenshot of the detector result with the date and the tool’s name. Scores can change between scans and between versions of a tool, so keep a record of what was actually shown.

Find out what you’re accused of

“The detector says it’s AI” can mean a tool wrote the whole piece, that it rewrote parts of it, or that your style changed from earlier work. Ask which of these is the concern, which passages were flagged, and whether anything besides the score is involved. Then read the policy itself. Some policies allow AI for brainstorming or grammar checks but not for drafting, so the exact wording matters.

Explain your process in writing

Write a short, plain account: how you chose the topic, where your sources came from, how many drafts there were and what changed between them. If you did use an AI tool for something, say what it did and what you wrote yourself. Specifics help you more than arguing that detectors are useless.

Ask how the decision gets made

You can reasonably ask whether the score is a screening flag or treated as proof, who reviews it, and how to appeal. If you’re confident in the work, offering to walk a reviewer through your drafts is a fair step.

Keep the exchange in email so there’s a record of what was said and when.