Plagiarism and AI at the Open University: automatic checks and the oral

The short answer

At the Open University every assignment is checked automatically, your tutor is not the one who sees the report, and an AI suspicion is settled by a live oral assessment rather than a score. That last point is the one to prepare for, because it is a conversation about your own work.

The governing document is the Academic Conduct Policy, version 1.2, effective January 2026, which replaced the July 2024 version. One detail worth knowing: it was approved by a delegate of the Director of Academic Services rather than by Senate — unusually low in the hierarchy for a UK academic-integrity policy, though it binds you all the same.

Everything is checked, automatically

Most universities leave the decision to a tutor. The OU does not:

“The files are automatically run through two types of similarity detection software soon after the published cut-off date. The Open University uses Turnitin to detect matches between your work and the module materials and other online sources, and to detect matches between assignments across current and historical cohorts of students.”

The scope is TMAs, emTMAs, EMAs and remote exams — in practice, everything you submit. The v1.2 change log confirms the tooling has been consolidated: “Removed references to Copycatch as we now use Turnitin for all forms of checking.”

Note the second half of that quote, because it has a consequence people do not think about. Your work is compared against historical cohorts, which means submissions are retained to be compared against. The OU does not publish a retention period, and no student opt-out from that comparison store is documented. If that matters to you, ask before you submit.

One practical implication: reusing your own earlier assignment is detectable by design, and self-plagiarism is a named category. If you want to build on something you already submitted, ask first.

Your tutor does not see the report

This surprises almost everyone, and it is genuinely reassuring once you understand it:

“Similarity detection software is checked by module teams in parallel with the marking by the tutor. Tutors do not normally have access to Turnitin reports, so if the module team finds evidence of academic misconduct in the software reports, this is not influenced by marks and your tutor may not be aware of it.”

Two things follow. Your mark is not quietly depressed by a similarity percentage — the person marking your work generally has not seen one. And the two processes are deliberately separated, so a high similarity score does not colour how your argument is read. If you have wondered whether a big block quote will cost you marks through the back door, at the OU the answer is that the marker is not looking at that number at all. What a similarity percentage does and does not mean is covered in what a Turnitin score actually means.

The oral assessment

This is the part of the OU’s regime that matters most, and it is the single clearest statement we have found at any university about how an AI allegation is actually resolved.

First, how suspicion arises — from a person, not a tool: “Tutors and markers may still be able to identify potential cases of academic misconduct when marking your work. They could identify possible changes in your writing style which may indicate that you have not written the assignment yourself.”

Then what happens next:

“If there is any doubt that your assignment is your own work, you may be asked to take an Additional Assessment which will be designed to test whether you have written it yourself. […] normally a live oral assessment, via video conferencing such as MS teams or Zoom, but occasionally a written task may be used instead. When the student is suspected of using GenAI, the assessment must be oral.”

Read that last sentence again. For AI specifically, the oral is not optional for the university — it must be oral. The OU has decided that the way to test authorship is to ask you about your work.

And declining has a price:

“If you do not engage with or agree to take the Additional Assessment offered then a decision will be taken without this piece of evidence. This could result in you receiving a zero for the assignment.”

So: take the oral. If you wrote the work, it is the best thing that can happen to you — far better than arguing about a number. Prepare the way you would for a viva: be able to say why you framed the question that way, what you read that changed your mind, which sources you rejected, what you would do differently. Bring drafts and version history if you have them. Our guides on proving you wrote it yourself and what to do when you are accused cover the preparation in detail.

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.

What counts as misconduct

The policy lists eight categories:

• Plagiarism
• Self-plagiarism
• Enabling plagiarism
• Collusion
• Contract cheating
• Unauthorised use of generative Artificial Intelligence and automated tools
• Falsification
• Misconduct in an exam

The AI category has three limbs, and the third is the one that catches honest students:

“Use of Generative AI without referencing how it was used, including as a starting point to generate reference material, to plan or produce content that appears in your assignment; Use of Generative AI or automated tools when this is not permitted in the module assessment guidance; Using Generative AI or other automated software tools to produce all or most of the content of your assignment, even when this is correctly referenced, unless the module permits this.”

That third limb is important. Referencing the tool does not make it acceptable to have it write most of your assignment. Disclosure fixes the honesty problem; it does not fix the authorship problem.

Paraphrasing and rewriting tools

We should be straight about this one, because we sell a rewriting product and the OU names the category directly. Its definition of an automated tool includes:

“those which reword or amend existing content such as translation tools, paraphrasing tools, LLMs, or automated re-writing tools.”

So at the Open University, a paraphrasing or rewriting tool sits in the same bucket as a large language model. Using one without acknowledging it is misconduct, and using one where your module does not permit it is misconduct even if you do acknowledge it.

If you are studying at the OU, that is a reason to be careful with any tool that rewrites your sentences for you — including ours. Our own check marks filler, hedging and tired transitions in your sentences and suggests plainer wording; nothing is rewritten for you. That is a deliberate design choice, and at the OU it is also the safer side of the line. But your module guidance is what governs, and if it says no automated tools, that includes the helpful ones.

What you must acknowledge

The OU specifies the form of the acknowledgement, which many universities do not:

“You must fully acknowledge any parts of your work where you have used Generative AI technology or automated software tools, including as a starting point or for reference, by specifying the tool and query-text used.”

“Use of Generative AI must be referenced using the guidelines for the referencing style specified in your module by specifying the AI tool used, the prompt-text (i.e., the question or instruction given to the AI tool) and the date the information was generated, as well as which parts of the assignment content were affected.”

Four elements, then: tool, prompt text, date, and which parts of the assignment were affected. The prompt text is the one people cannot reconstruct afterwards — so if you use anything, save the conversation as you go. Our AI disclosure statement guide has wording that covers all four.

Penalties

The OU does not publish a numeric tariff table in this policy. The scale it does set out runs:

• An informal caution or warning, where an offence is judged not serious enough for a disciplinary penalty. It still goes on your student record and “can be taken into account if you are referred for an academic conduct investigation in future”.
• A disciplinary penalty, which “will include a mark deduction which may mean your score is capped at the pass mark for your module or reduced to zero”. Note the sting in the tail: if the plagiarism is extensive, “the mark you are given may still be lower than the cap”. A note of it “will stay on your student record for the remainder of your studies”.
• Referral to the Central Disciplinary Committee, which handles serious and repeat offences and “can also impose a further range of penalties according to severity”.

Exclusion and degree revocation are not enumerated in the Academic Conduct Policy itself — they sit in the separate Code of Practice for Student Discipline, which we have not extracted here. We would rather tell you where to look than guess at the list.

One extra exposure if you are on a professional programme: “any case of academic misconduct confirmed with the application of a penalty may be reported to the relevant professional body […] In some cases, consideration will also be given to whether a Fitness to Practise investigation should be undertaken.” For nursing, social work, teaching and similar, the academic consequence is not the only consequence.

Appeals are limited to three grounds — procedural error, genuinely new evidence, or a penalty that is unreasonable in the circumstances — and must be made within 28 days of the outcome letter. “You cannot appeal an outcome just because you disagree with it.”

It can happen after you graduate

Stated plainly in the policy, and rarely stated anywhere else:

“This can take place months or even years after the offence was originally committed, for instance if your assignment material is found on an external website some time after it was submitted. You can still be investigated for academic misconduct after you graduate from The Open University.”

The example given is worth noticing: assignment material appearing on an external website. If you have ever uploaded your own work to a note-sharing or essay site, or to a checking service that stores what you give it, that copy can surface later and be matched against you. Before you put a draft into any online service, find out whether it keeps it.

Before you submit

• Assume it is checked. At the OU that is not a risk assessment, it is the documented process.
• Read your module assessment guidance on AI. Permission is set per module, and the policy defers to it.
• Save prompts and dates as you go. The required acknowledgement needs the prompt text, which you cannot recall later.
• Keep drafts and version history. The route out of an allegation is an oral about your own work.
• Be careful where you upload a draft. Comparison against historical cohorts and external websites is explicitly part of the process.

If you want a read on your own text before you submit, our AI check is free up to 1,500 words, with no account and no name. Two things we always say: our published error rates are measured on English and German corpora, and no detector — ours included — produces proof. At the OU that second point is more than a disclaimer, because the university has built its process around an oral rather than a score.

Sources

• The Open University: Academic Conduct Policy — version 1.2, effective January 2026, approved by a delegate of the Director, Academic Services.
• The Open University: Plagiarism and academic conduct — policy hub, including the policy summary.
• The Open University: Code of Practice for Student Discipline — where the full penalty list sits.

This page is orientation, not legal advice. What binds you is your module assessment guidance and the OU’s own policies.