The useful result
Assemble an evidence folder and a short, calm reply for a flagged piece of work.
- Assemble an evidence folder and a short, calm reply for a flagged piece of work.
You wrote the essay yourself and a detector says it is AI-generated. The shock is real. A Stanford study found that seven detectors flagged, on average, more than 61 percent of TOEFL essays by non-native English writers as AI-generated, and OpenAI’s own help centre for educators says detectors have not proved reliable. A score alone is a weak basis for a serious accusation.
Still, you need a response that is calm and backed by evidence. This lesson helps you collect the record of how you wrote the work, write a short factual reply and follow your institution’s process. It does not help anyone hide AI use; it helps honest writers show their work.
The examples below are illustrative exercises, not recorded results from any tool. Sources were checked on 24 October 2026; menus, limits and prices change, so confirm details on the linked pages.
Before you start
- The flagged work and any message about the flag
- Your institution’s academic integrity policy
- Access to your document history, notes and drafts
Review evidence
Documentation-based review · 2026-10-24
Primary documentation and official guidance checked on 24 October 2026. Examples, checks and exercises are original illustrations, not recorded tool tests. Cover artwork is a conceptual illustration.
Pause, keep your work as it is and ask for the details
Do not rewrite the piece to “lower the score”; that can destroy your evidence. Ask politely for the detector report, the policy that applies and how the decision will be made. Read your institution’s appeal process. Keep every message.
Check your result
You have asked for the report and the process, and you have not altered the flagged file.
Gather the record of how you wrote it
Collect version history from your document tool with its timestamps, notes, outlines, research links, earlier drafts, search history and any feedback from a tutor. Export or screenshot them with dates. A steady build-up of changes over days is much stronger than a final file alone.
I am writing a short, calm reply to a teacher who flagged my work using an AI detector. Facts: [what you wrote, when, how, evidence you have]. Write a reply of under 150 words that: states I wrote the work myself, lists the evidence I can share, asks to discuss it and asks for the detector report and policy. Polite, factual, no accusations, no anger. Do not add any facts I did not give.Example / expected result
Illustrative example; not a recorded model result.
“I can share the document history showing my drafts from 3 to 9 October, my research notes and my outline.”
Check your result
Your folder shows dated drafts, notes and sources that lead to the final text.
Send the reply, ask for a meeting and follow the process
Send your reply through the official channel, offer to walk through your drafting process in person and follow each step of the appeal process. In the meeting, show the history and explain your choices. You may mention that detectors have known error rates, but let the evidence lead.
Check your result
Your reply is sent, a meeting is requested and you have the evidence ready to show.
When it goes wrong
If you did use AI in a way the rules prohibit, honesty is better than a defence you cannot support. Version history may not exist if you typed in an app without it; start using a tool with history now. Policies and appeal rights vary by school and country.
Make it your own
Start your next assignment in a document with version history, save notes in the same place, and practise exporting the history so you could share it quickly.
Useful next resources
Sources and review
- Stanford HAI: AI detectors are biased against non-native English writers
- OpenAI Help: how educators can respond to AI-generated student work
Last checked Oct 24, 2026. Features and account limits can change. Verify important results against your original material.
Did this help you complete your task?
Tell us what needs improving. No account or private output required.
Report incorrect information or a broken linkGenius Forges Editorial
Practical AI lessons with checked sources, labelled examples and human verification.
