Falsely accused of using AI?
What to do —
and how to show the work is yours.

Being accused of something you didn't do is horrible, and AI detectors get it wrong more often than universities admit. If this has happened to you, take a breath: a detector flag is a probability, not proof. Here's how to respond calmly and put your case together.

In this article
  1. A detector flag is a probability, not proof
  2. Stay calm and read your university's process
  3. Gather your evidence
  4. How to respond to the allegation
  5. Why this keeps happening
  6. How to protect yourself next time

A detector flag is a probability, not proof

Start here, because it changes everything about how you respond. An AI detector does not "catch" AI — it outputs a likelihood score based on statistical patterns, and those scores are far less reliable than the headline numbers suggest. Independent testing has put some detectors' real-world accuracy as low as 22–40%, and a widely cited Stanford study found detectors wrongly flagged around 61% of essays by non-native English speakers as AI-written.

This is now well enough established that institutions are backing away from these tools: the University of Waterloo discontinued Turnitin's AI detector in September 2025, and Curtin University followed in January 2026, both citing unreliability and the risk of wrongly accusing students. In July 2026, HEPI described AI detection in UK universities as a "fairness crisis." So if the case against you rests on a detector percentage, that percentage is contestable — and you are not the first person to contest it.

The key point

A number from an AI detector is evidence to be weighed, not a verdict. The burden is on the university to show misconduct on the balance of probabilities — a single tool's score usually isn't enough on its own.

Stay calm and read your university's process

An allegation is not a finding. UK universities have a formal academic misconduct procedure, and you have rights within it: to be told clearly what you're accused of and which policy applies, to see the evidence, and to a genuine chance to respond before any decision is made.

So before you reply to anything:

You don't have to face it alone

Your students' union adviser, personal tutor and student support services exist for exactly this. Being accused is stressful — lean on them, and look after yourself while it's ongoing.

Gather your evidence

The single most powerful thing you can show is your writing process — because a genuine piece of work leaves a trail, and generated text usually doesn't. Pull together whatever you have:

If you can talk a marker through how you built the argument — why you cut a paragraph, where a source changed your mind — you're demonstrating the one thing an AI can't fake: your own reasoning.

Talk it through with Sage.
Sage can help you explain your own argument and reconstruct how you built the essay — useful preparation before a meeting. Free to start.
Ask Sage →

How to respond to the allegation

When you put your response together, keep it factual and calm:

A note, not legal advice

This is general information, not legal advice, and every university's process differs. For serious cases or appeals, your students' union or a qualified adviser can help you prepare formal representations.

Why this keeps happening

It helps to know you're not imagining the unfairness. Detectors are trained to spot "average", predictable prose — which means clear, well-structured, grammatically clean writing can look "AI" to them. Students who write in a tidy, formulaic way, who use tools like Grammarly heavily, who are neurodivergent, or who speak English as an additional language are all flagged more often, through no fault of their own. The tool's weakness becomes the student's problem. That's precisely why sector bodies are now urging universities not to rely on detector scores as evidence.

How to protect yourself next time

You can't control a detector, but you can make your own process easy to show — and see problems before a marker does. A few habits help:

Being accused when you did nothing wrong is not something you should have to manage. But knowing how detectors actually work, keeping your process, and checking your own writing first put you in the strongest possible position if it ever happens.

See how your own writing
reads — before anyone else does.
A private read on your writing patterns, an AI-risk indicator, citation checks and a grade estimate. Stored on UK servers, never shared with your university. Free to start — no account needed for your first scan.
Check my essay free →