Wrongly flagged by an AI detector? Here's what to do

For students and parents. Written by a company that makes an AI detector, which is exactly why we publish this page: every detector, including ours, sometimes flags work a human wrote. If that has happened to you, the process below protects you better than panic does.

First: what a detector score actually is

A detector score is a probability, not a verdict. "78% likely AI" does not mean "this student cheated"; it means a statistical model found patterns that resemble AI output. Real human writing produces those patterns too. On our own published measurements, this detector wrongly flagged 4 of 400 real student essays and 5 of 517 professional pieces - and published research has found some detectors flag essays by non-native English speakers at far higher rates. The full numbers, including what we miss, are on our methodology page.

Exam-board guidance points the same way. JCQ, which oversees GCSEs and A-levels, treats a detector result as one piece of evidence inside a judgement made by a teacher who knows the student, never as proof on its own. If your work was flagged, the score is where the conversation starts, not where it ends.

Gather your evidence - you probably have more than you think

  • Version history. Google Docs and Word both keep it. A document that grew over hours and days, with edits and deletions, is strong evidence of human writing. Open it before your meeting, not during.
  • Drafts, plans and notes. Outlines, rough drafts, research notes, photos of handwritten planning - anything that shows the work developing.
  • Your other writing. Earlier essays and classwork in your own voice. Consistency with how you normally write is evidence in your favour.
  • Browser or account history showing research on the topic, if you are comfortable sharing it.

What you are entitled to ask for

These are reasonable questions in any school or university process, and they are the ones a fair process should already answer:

  • Which passages were flagged, and why? A bare percentage is not something you can respond to. Specific passages are.
  • What are the tool's published error rates? Every detector has false positives. If the vendor publishes its rates, they can be weighed; if it publishes nothing, that is worth saying out loud.
  • What other evidence is being considered? Your drafts, your classwork, a conversation about the essay's content. A decision resting on a score alone is not a safe decision.
  • Who makes the decision, and how do I appeal? The decision should sit with a person, not a tool, and there should be a named route to challenge it. Ask for both in writing.

One practical tip: offer to talk about the work. A short conversation about what you wrote, why you structured it that way, and what you found hard is often the fastest way to show it is yours.

If the flag came from Is It AI?

Contact us at hello@isitai.co.uk with the text and the result. We log challenged results, review them, and use them to refine the system - and our results are designed to be read as a screening signal for a teacher to weigh, never as proof. If a result of ours is being used as the sole basis for a penalty, you may point the decision-maker at this page and at our methodology, which say so in plain terms.

Read more