Glyphtell Research Prototype

Caught using AI on an assignment: what to do next

If you did use AI and it has been noticed, the next few days matter more than the mistake. This is a plain guide to getting through it with the least lasting damage.

Last reviewed 12 November 2026

Most of this site is about people who are wrongly flagged. This page is for the other situation: you used an AI tool in a way your course did not allow, and your instructor has raised it. You are not the first, the process is survivable, and how you handle it now has a large effect on the outcome.

First 24 hours

  1. Do not reply in a panic. Read the message twice. Note exactly what is being alleged, about which piece of work, and what you are being asked to do by when.
  2. Do not delete or edit anything. Leave the document, its version history and any chat logs as they are. Altering evidence turns a first offence into something much worse, and deletions are often recoverable.
  3. Read the actual rules. Find the course syllabus and your institution's academic integrity policy. You need to know what was permitted, what the process is, and what the range of outcomes looks like.
  4. Work out precisely what you did. "I used AI" covers everything from fixing grammar to submitting generated text unchanged. Write down, for yourself, which parts of the work came from where.
  5. Find your support service. Many colleges have a students' union adviser, an ombuds office or an academic advice service that helps with exactly this, confidentially and for free.

Be honest, and be exact

In most integrity processes, an early and accurate account is treated far more leniently than a denial that later collapses. If a hearing panel concludes you lied to it, that often becomes a second and more serious problem than the original one.

Being honest does not mean agreeing to everything. Admit what is true and be clear about what is not. If you used a chatbot to outline but wrote every sentence yourself, say that. If two paragraphs were generated and the rest was yours, say which two. Our guide on whether to admit using AI goes through this decision in detail.

Do not try to fix it with more tools. Rewriting the work, running it through a paraphraser or a "humanizer", or producing drafts after the fact are all easy to spot in a version history and will be read as an attempt to deceive.

What to bring to the meeting

What usually happens

Outcomes vary a great deal between institutions and depend on your history. For a first incident, common results are a warning, a reduced or zero mark for the assignment, a required resubmission, or an integrity course. More serious or repeated cases can lead to failing the module, suspension or, rarely, expulsion. See what the consequences can be for the questions to ask about records and transcripts.

Afterwards

If you did not use AI, this is the wrong page. Start with Accused of using AI when you didn't.

This page is general information, not legal advice. Procedures and penalties differ between institutions; your own institution's written policy is the authority.