Plagiarism and AI at Manchester Met: at level 7, a first minor offence is already a zero

The short answer

Manchester Met’s penalties scale with how far into your degree you are, and at level 7 — Master’s study — even a first, minor offence is Category B: a written warning and a mark of zero for the assessment. The same act by a first-year would be Category A, remarked with the offending elements removed and no cap.

That is unusual, and it is published, which is more than most UK universities do. The governing document is the Academic Misconduct Policy 2025/26, version 3.2, approved by Academic Board, carrying an approval date of March 2025. If you are writing a Master’s dissertation at Manchester Met, the matrix below is the single most important thing on this page.

First: this is not the University of Manchester

Manchester Metropolitan University and the University of Manchester are two separate institutions, a mile apart, with entirely separate academic misconduct regimes. Students confuse them constantly, and the confusion is expensive here because the two universities have made genuinely different choices — including on AI detection, where the University of Manchester has published a position and Manchester Met has not.

If you are actually at the University of Manchester, our page on plagiarism and AI at the University of Manchester is the one you want. If your student email ends in @stu.mmu.ac.uk and your policies live at mmu.ac.uk, stay here.

One more thing to check before you rely on any version of the document. Manchester Met’s student page links a policy it labels “Academic Misconduct Policy 26/27” at a URL whose slug reads misconduct-policy-24-25. We record that as found rather than guessing which is right: on that site the slug is not a reliable signal of which edition you are opening. Check the front matter inside the document — the version number and the approval date — rather than the address bar.

The penalty matrix, and why level 7 is different

Manchester Met sets the outcome by combining two things: your level of study, and the seriousness and history of the offence. Here is the published matrix.

Read down the level 7 row and the principle is explicit: the same act is punished harder the further into your studies it happens. The reasoning is not hard to reconstruct — a Master’s student is not learning to reference for the first time — but few universities put it in a table.

For a taught postgraduate the practical consequences are these. There is no minor-offence landing zone. The Category A outcome, where the work is simply remarked with the offending elements removed, is not available to you. A first case, however small, means zero for that assessment. And a second offence at level 7 is Category C — course failure.

Where reassessment is allowed, it is “capped at module pass mark”. So even a recovered outcome costs you the grade.

One more thing that matters at this level: Manchester Met’s plagiarism definition carries no minimum threshold. There is no published quantity of unattributed material below which nothing happens. That does not mean every small slip becomes a case — the categories still distinguish minor from major — but it does mean you cannot reason from a percentage.

The three penalty categories

The matrix outputs one of three categories, each defined in the policy.

• Category A (minor) — the coursework is remarked with the misconduct elements removed; exams are unchanged; there is no cap on the resulting mark. The case is recorded, a written warning is issued, and you are required to engage with developmental support.
• Category B (major) — a written warning and a mark of 0 for the assessment, recorded on your file.
• Category C (severe) — the student is deemed to have failed the course, recorded on your file.

Category A deserves a note, because it is genuinely lenient by UK standards: the work is remarked without the offending material rather than being zeroed or capped. If your case is Category A, engage with the developmental requirement properly. It is the outcome that keeps your mark intact, and it is the outcome that is off the table if there is ever a second case.

Every category is recorded, including A. That record is what turns the next incident into a “2nd offence” row of the matrix, which at every level is a harder outcome.

It is worth understanding how the matrix and the categories interact, because students often read one without the other. The matrix does not decide whether misconduct occurred; it decides what happens once a finding has been made. The inputs are your level of study, whether the case is classed minor, major or severe, and whether it is your first. The output is a letter, and the letter is what the penalty means.

Two consequences follow. First, an argument about seriousness is worth making explicitly: minor versus major is the difference between A and B at levels 5–6, and between an intact mark and a zero. Second, the “severe” column bypasses the first-offence logic entirely — a severe case is Category C at every level, including a first-year’s first offence.

Manchester Met does not publish what separates minor from major from severe in quantified terms, so do not try to reason from the size of the passage or the percentage of the document. If your case is being classed at a level you think is wrong, that is a point to make in the investigation, in your own words, with your evidence.

Unauthorised generative AI use, defined

Manchester Met gives AI its own misconduct category rather than folding it into plagiarism. The policy definition:

“Unauthorised generative AI use — Where a student has made unauthorised use of generative artificial intelligence within their assessment and has attempted to gain an unfair advantage by passing this work off as their own.”

The student-facing page uses the same phrase — “unauthorised use of generative artificial intelligence (AI)” — so this is consistent across both surfaces.

The load-bearing word is unauthorised. Manchester Met is not saying AI is forbidden; it is saying that use outside what your assessment authorises is an offence. Which means the university-level policy cannot answer the question you actually have. The authorisation comes from your assessment brief, your module handbook and your supervisor. If none of them say, ask — in writing, and keep the reply.

The definition also contains a second element that is worth noticing: “has attempted to gain an unfair advantage by passing this work off as their own”. That is the same test as the general definition of misconduct on the student page, which is “any action that could give you an unfair advantage in coursework, exams or any other assessed work”. The corresponding definition of academic integrity is “being honest and transparent, and demonstrating rigour and accuracy in your work”.

Manchester Met links guidance titled “Referencing generative AI in your assessments”, which tells you that acknowledgement is expected where use is authorised. What it does not do is publish a prescribed form of words. So keep your declaration short and specific — which tool, for what, on which parts — and check whether your course wants something particular. Our AI disclosure statement guide sets out that shape.

The other named offences

The assessment regulations page enumerates the categories in one sentence:

“Academic misconduct exists in numerous forms, including collusion, contract cheating, cheating, exam disruption, falsification of data, plagiarism and self-plagiarism.”

Contract cheating gets its own definition in the policy, and it is broader than the phrase suggests:

“Contract cheating — When a third party completes work for a student who then submits it as their own” — including friends or family, with or without payment.

That catches something people genuinely do not think of as cheating: a partner, a housemate or a relative writing a section for you. No money need change hands, and no commercial service need be involved.

Self-plagiarism is on the list too — reusing your own previously assessed work. If you are building a dissertation on a topic you have written about before in the same programme, raise it with your supervisor rather than assuming your own words are free.

If you are writing a Master’s dissertation

Almost everything on this page compresses into one paragraph for a taught postgraduate, so here it is on its own.

Your dissertation is a level 7 assessment. Any finding of academic misconduct on it is at least Category B, which is zero for the assessment. There is no first-time discount at your level and no minor-offence route to a remark. A second finding is Category C — deemed to have failed the course. And a reassessment, where one is allowed at all, is capped at the module pass mark.

That is a harsh structure, but it is also a legible one, and legibility is useful. It tells you exactly where to spend your care. Three places, in order.

• Referencing, mechanically. With no minimum threshold published, the safe practice is that every borrowed idea, sentence and figure carries an attribution — including the ones you paraphrased so heavily that they feel like yours. Build the bibliography as you write, not at the end, because the end is where citations go missing.
• The AI authorisation. Ask the module leader or your supervisor, in writing, what is authorised for the dissertation specifically. A dissertation is not covered by whatever your taught modules allowed — it is a separate assessment with its own brief.
• Anyone who touches the text. A proofreader, a friend who “tidied a chapter”, a study group that drafted a shared literature review — contract cheating at Manchester Met covers third parties with or without payment, and collusion covers unauthorised collaboration. Get the boundary confirmed before someone helps, not after.

And keep the record. Over a dissertation-length project you will produce months of drafts almost incidentally; the only mistake is overwriting them. Work in a file with version history, or save dated copies at each milestone. It costs nothing until the day it is the only thing you have.

The standard of proof, published in plain words

Manchester Met publishes the test it applies, which is unusual — of the UK institutions we looked at for this batch, it was the only one where the standard of proof appeared on a primary source in ordinary language:

“The University is satisfied that, on the basis of the evidence available, that academic misconduct is likely to have occurred.”

The grammar is awkward — the doubled “that” is in the original — but the meaning is clear. That is the balance of probabilities, expressed in words rather than by name: more likely than not, on the evidence available.

It is worth being realistic about what that means if you are answering an allegation. You are not facing a criminal standard, and the university does not have to be certain. What your evidence has to do is make the university’s account less likely than yours — which is a job for the record of how the work was made, not for an argument about how confident anyone should be.

In practice that means drafts with dates, version history, your reading notes, your search history in the library catalogue, the emails to your supervisor. We set out what a persuasive authorship record looks like in how to prove you wrote it yourself, and the order to do things in if a case has already been opened in what to do when you are accused of using AI.

If a detector score is what started this, our guide to an AI detector flagging your thesis sets out what that number does and does not establish on its own.

Investigation, appeal, review, OIA

Manchester Met publishes the whole route, and the roles at each stage.

• Minor and major cases — the Head of Department, or a nominee, acts as Investigating Officer.
• Severe cases — an Assessment Disciplinary Panel, whose default outcome is course failure. If you are called to one of these, understand that the starting point is Category C, and prepare accordingly.
• Penalties are applied centrally by Assessment Management, not by the department. The decision and its execution are separated.
• Appeal within 10 working days, on three grounds: that the decision “was unreasonable”; that there was a material or procedural irregularity prejudicing the case; or that new evidence emerged after the investigation.
• The Appeal Panel is a Senior Officer as Chair, two uninvolved senior academics, and a Student Panel Member.
• A Review stage is available within 14 days of the appeal outcome.
• The OIA — on conclusion you receive a Completion of Procedures letter and may take the case to the Office of the Independent Adjudicator within 12 months.

Two things to act on. Ten working days is short. It is roughly two weeks, and it begins when the decision is issued, not when you feel ready. Note the date the moment a decision arrives.

And the three appeal grounds are exhaustive. An appeal that says only “the penalty is too harsh” fits none of them; an appeal that says “the penalty applied Category B when the matrix gives Category A for my level” is a decision that was unreasonable, and an appeal that says “I was not given the chance to respond before the finding” is a procedural irregularity. Write to the grounds.

The composition of the Appeal Panel is worth knowing too, because it tells you who you are writing for. A Senior Officer chairs; two senior academics who had nothing to do with the case sit with them; and there is a Student Panel Member. None of them was in the original investigation. So do not assume any background knowledge — set out what the assessment was, what you did, what was decided, and why one of the three grounds is met, in that order, plainly.

The Review stage that follows is not a second appeal on the merits; it is a further step available within 14 days of the appeal outcome. Note that date the way you noted the first one. And when the internal process ends, the Completion of Procedures letter is not merely a formality — it is the document the Office of the Independent Adjudicator requires, and it starts the twelve-month window for taking the case externally. Diarise its date on the day it arrives.

What Manchester Met does not publish

The tariff and the process are well documented. The detection side is not documented at all. On the public web we could not establish:

• Which similarity tool is used. Neither the assessment regulations page nor the academic integrity page names Turnitin or any other product; the assessments page does not mention Turnitin at all.
• Whether AI-writing detection is used, or any position on it. The policy penalises unauthorised AI use without saying how it is identified. Treat this as not published rather than as an answer.
• Whether any student self-check route exists.
• Whether submissions are retained in a comparison repository, and whether any opt-out exists.
• A prescribed AI declaration wording. Acknowledgement is clearly expected where use is authorised, but the form of words is not published.
• Any power to revoke a degree after graduation.
• Any annual case counts.

Some of that is probably documented where you can see it and we cannot — module handbooks, the virtual learning environment, the guidance page on referencing generative AI. If a point on that list matters to you, ask your course team rather than assuming the silence is the answer. The questions worth putting in an email are short: is this assessment checked by any similarity or AI tool; if a report is generated, will I be shown it; and is generative AI permitted for this assessment and in what scope.

Because no detection position is published, there is no honest way to tell you what a checker will or will not do for your position at Manchester Met. There would not be a good answer anyway: our own published error rates are measured on English and German corpora, and no detector — ours included — produces proof of who wrote a document. If you want the full argument, see how accurate AI detectors really are, and for the number most students actually meet, what a Turnitin score actually means.

Before you submit

• Know your level. If you are at level 7, there is no minor-offence outcome available to you. Work as though a first case costs you the assessment, because it does.
• Get the AI authorisation in writing. The offence is unauthorised use, and only your assessment brief or module leader can tell you what is authorised.
• Acknowledge authorised use — which tool, for what, on which parts.
• Do not let anyone else write any of it, paid or unpaid. Manchester Met’s contract cheating definition covers friends and family.
• Check before reusing your own earlier work. Self-plagiarism is a named category.
• Keep dated drafts. With the standard set at “likely to have occurred”, your working record is what shifts the balance.
• Open the actual policy document and read its front matter rather than trusting the URL. The slugs on that site do not track the editions.

Sources

• Manchester Metropolitan University: Academic Misconduct Policy 2025/26 — version 3.2, approved by Academic Board, approval date March 2025. The category definitions, the matrix, the standard of proof and the process.
• Manchester Metropolitan University: the edition the student page labels “Academic Misconduct Policy 26/27” — the label and the slug disagree; check the document’s own front matter.
• Manchester Metropolitan University: Academic integrity (student page) — the definitions of integrity and misconduct, and the referencing-AI guidance link. The page carries no date.
• Manchester Metropolitan University: Assessments — the enumeration of misconduct categories.

For comparison across UK institutions, plagiarism and AI at Leeds covers a university with a published red/amber/green AI scheme and a strict-liability clause, and plagiarism and AI at Sheffield covers one that has published an explicit decision not to use AI detection tools. If you want to see what our own check reports on a document, it is at the AI check.

This page is orientation, not legal advice. What binds you is your assessment brief and Manchester Met’s Academic Misconduct Policy in the edition current for your year of study.