ESL Humanizer

How to appeal an AI-cheating accusation (with a free letter template)

By ESL Humanizer editorial teamUpdated 7 min read

Short answer

A strong appeal is short, factual and evidence-led: state that the work is your own, name the tool and score that flagged it, explain your writing process, attach dated evidence such as version history and drafts, cite research on detector false positives, and ask for a meeting or an oral review of your work.

If you have been told your work was AI-generated and you wrote it yourself, a clear written appeal is often the most important thing you will send. This guide gives you a template you can copy, explains each part, and lists the evidence that makes an appeal hard to dismiss.

If you have not yet talked to your instructor, start with our step-by-step plan — many cases are resolved informally before a formal appeal is needed.

Before you write

  • Find the formal process and deadline in your institution's academic integrity policy. Missing a deadline can end an appeal before it starts.
  • Collect your evidence first — version history, drafts, notes, sources. Write the letter around what you can show.
  • Ask for help. Student advice services, international student offices and writing centers regularly help with appeals, and some can review your letter.

What a strong appeal includes

  1. A clear statement that the work is your own and that you are appealing the finding.
  2. The specifics of the accusation — tool, score, flagged passages — or a request for them.
  3. Your writing process, with dates, told briefly.
  4. A list of attached evidence.
  5. Why the detector result is unreliable, backed by research rather than frustration.
  6. A specific request — a review, a meeting, an oral examination of the work.

The appeal letter template

Replace everything in [square brackets]. Keep it to one page if you can; your evidence does the heavy lifting.

Subject: Appeal regarding AI-use finding — [Course code], [Assignment title]

Dear [Name / Academic Integrity Panel],

I am writing to appeal the finding that my [assignment title], submitted on [date] for [course code], was generated with AI. I wrote this work myself, and I would like the opportunity to demonstrate how I produced it.

1. What was flagged
I understand the work was flagged by [tool name] with a score of [score/percentage], and that [specific sections] were highlighted. [If you do not know this, write: "I have not yet been told which tool or score was used, and I respectfully request that information."]

2. How I wrote the assignment
I began researching on [date], created an outline on [date], and wrote the draft over [number] sessions between [date] and [date]. I revised it after [feedback / writing center session / re-reading]. [If you used any tools, list them honestly, e.g. "I used the spell-checker in Word and an online dictionary for individual words."]

3. Evidence of authorship
I have attached:
- The version history of my document, showing it being written and revised over time
- My outline and notes
- [Earlier drafts / research notes / reference list / emails]
- [Previous assignments that show my usual writing style]

4. Why a detector score is unreliable in my case
English is not my first language. A peer-reviewed Stanford study (Liang et al., 2023, Patterns) found that seven AI detectors misclassified an average of 61% of TOEFL essays written by non-native English speakers as AI-generated, while classifying native speakers' essays almost perfectly. OpenAI withdrew its own AI-text classifier in 2023 because of its low accuracy. I believe my writing was flagged because of the simpler vocabulary and sentence structures typical of second-language writing, not because I used AI.

5. What I am asking for
I respectfully ask that the finding be reviewed in light of this evidence. I would welcome a meeting to discuss my work, and I am happy to explain my argument and sources orally or to write a short passage under supervision.

Thank you for your time and consideration.

Sincerely,
[Full name]
[Student ID]
[Program and year]
[Email]

What to attach

EvidenceWhy it helps
Version history (Google Docs, Word/OneDrive)Shows the document being built and revised over time
Outline, notes, earlier draftsShows your thinking developing before the final text
Research trail and reference listShows you read and used the sources you cite
Previous graded workShows the flagged work matches your usual style and level
Research on detector biasShows the score is weak evidence for non-native writers (Liang et al., 2023)

Useful, verifiable facts to reference: the Stanford study's 61% average misclassification of TOEFL essays (Liang et al., 2023); OpenAI withdrawing its classifier over low accuracy (OpenAI); and Vanderbilt disabling Turnitin's AI detector over false positives (Vanderbilt). The background is explained in why AI detectors flag human writing.

Mistakes that weaken appeals

  • Attacking your instructor. Criticize the tool, not the person.
  • Hiding tools you used. If a grammar checker or translation dictionary touched your work, say so. Being caught out later is far worse.
  • Relying on another detector's “human” score. Detectors disagree with each other, so a clean result elsewhere proves little. Process evidence is stronger.
  • Writing too much. A focused page plus organized attachments beats five pages of argument.

Polish your appeal's English — without changing what you say

Paste a paragraph of your letter. We fix grammar and phrasing, keep your meaning, and highlight every change.

Try it free

Quick answers

Should I send the appeal by email or through a portal?

Use whatever channel your policy specifies. If it allows email, keep a copy and ask for confirmation of receipt.

Can I ask for an oral examination of my work?

Yes — offering to explain your work or write under supervision is one of the most persuasive things you can do, because it tests authorship directly.

Is it okay if my letter isn't in perfect English?

Yes. Clarity matters more than style. Short sentences and a clear structure are perfectly professional.

Sources

  1. Liang, Yuksekgonul, Mao, Wu & Zou — Patterns (Cell Press) (2023). GPT detectors are biased against non-native English writers
  2. OpenAI (2023). New AI classifier for indicating AI-written text
  3. Vanderbilt University (2023). Guidance on AI Detection and Why We're Disabling Turnitin's AI Detector