AI Detection at UK Universities: Which Tools They Use and What the Policies Say
Which UK universities use Turnitin's AI writing indicator and which don't, what the Russell Group principles and coursework rules say, and how OIA and SPSO complaints work.
The AI detection UK universities use, where they use any, is Turnitin's AI writing indicator. Several large universities now say on their own pages that they don't use it. Imperial College London, the University of Glasgow, Newcastle University and the University of Westminster all state this. The University of Bristol keeps the reports on and tells staff to treat a score as supporting evidence only.
Two decisions shape what you face as a student. Your university decides whether the Turnitin indicator is switched on, and it sets the rules on how students may use AI in coursework. If a concern is raised, the academic misconduct procedure follows, and in England, Wales and Scotland an independent body can review how the university handled your case.
What AI detection UK universities use in 2026
Turnitin's indicator is the one that the AI detection UK universities discuss in their guidance. Jisc's National Centre for AI in September 2023 said that it wasn't aware of any UK university or college having formally adopted any AI detection other than Turnitin. Jisc is a body that advises UK colleges and universities on digital tools. Its members had asked it about alternatives to Turnitin.
The indicator is a setting inside Turnitin. From the admin settings page, Turnitin's administrators can enable or disable the AI writing feature. When this feature is disabled, the similarity report continues to work normally. A university can use Turnitin to check similarity and show no AI scores at all. Our list of universities that stopped using Turnitin's AI detector covers the US cases.
Students can't see Turnitin's AI writing report. Turnitin's help center states that the reports are not visible to students. The student who wants to know their score must request it from the instructor, who will then be able to download the report and share it with them. According to the guide to a Turnitin AI score not showing, the missing number on your submission page is perfectly normal.
Not all UK universities are sticking with Turnitin. According to an August 2026 update to the University of Westminster's staff FAQ, its executive board chose not to renew its Turnitin contract, which runs out on 31 July 2027. From Semester 2 of 2026/27, new assignments will go onto Blackboard Assignment with SafeAssign (Blackboard's similarity checker). Westminster notes that SafeAssign lacks an AI writing detection feature, as does Turnitin at Westminster.
UK universities that don't use Turnitin's AI indicator
Several UK universities state on their own pages that they don't use Turnitin's AI indicator or any AI detector. The reasons they give are accuracy and fairness, and Glasgow adds data protection when student work goes to outside tools.
| University | What its own page says | Page date |
|---|---|---|
| Imperial College London | Does not use Turnitin's AI writing detection tool in its assessment processes; staff decide on academic judgment and evidence | 2026-27 academic year |
| University of Glasgow | Does not employ AI detection tools; staff should not upload student work to AI detection platforms | GenAI Guidance 2026 |
| Newcastle University | Does not use the AI detection feature within Turnitin; may use an AI detection tool within a misconduct investigation if required | March 2025 |
| University of Westminster | Turnitin at Westminster has no AI writing detection feature, and neither does SafeAssign; its approach does not rely on automated detection | August 2026 |
| University of Cambridge | Advises staff not to rely on AI detection software, which it says is not proven accurate or reliable | Current staff guidance |
Glasgow's guidance is the most detailed of these pages. It says academic judgment, based on knowing a student's abilities and how AI-generated text tends to read, is a more reliable basis than software. It lists signs that may justify a closer look, such as fabricated references, references that don't support the claim, generic conclusions and a marked change from the student's earlier work. One sign is specific to the UK: unexpected Americanisms in work otherwise written in British English.
Glasgow also says these signs do not automatically mean a student used AI. Before a formal referral, staff may hold an exploratory interview, ask for a written statement or ask to see earlier drafts.
Newcastle's wording leaves room for a tool inside an investigation. It does not use Turnitin's indicator on submissions, and it says it will only use an AI detection tool as part of an academic misconduct investigation if required. Its library also strongly advises students against using AI detectors on their own work and calls them highly unreliable.
UK universities that keep the indicator on, and the limits they set
Some UK universities do use the indicator, and their guidance limits how staff read it. In November 2025, the University of Bristol's Digital Education Office wrote: Turnitin has updated the AI writing report on 14 October 2025. This post says that sometimes these reports can generate false positives, and it's important to check them with subject experts and use them as a complement to academic judgment.
Two universities that switched it on in 2023 wrote similar limits. Ulster University enabled the indicator after a working group review, on the condition that guidance explained the tool's limits. Its July 2023 page says a score will not definitively prove that work was plagiarised. Aberystwyth University's digital education team wrote in May 2023 that staff could see the AI score and students could not. It advised against bringing a case when the score was the only suspicious sign.
Turnitin gives the same warning about its own score. It says the model may not always be accurate and that the score shouldn't be the only basis for action against a student. The published figures are covered in the post on Turnitin's false positive rate.
Why many UK universities opted out in 2023
Turnitin launched its AI writing detection on 4 April 2023 and gave universities little time to decide. Bournemouth University wrote that the feature would be used from early April for every Turnitin customer unless the institution opted out. The University of Manchester opted out of the first release and said it did so along with the majority of UK universities.
Bournemouth listed its reasons in April 2023. There was no chance to preview or test the feature. The university's academic offenses policy was still being updated to take account of AI. There were no clear instructions from Turnitin about how the feature worked or its reliability. Bournemouth said that opting out meant that the similarity report would remain unaffected; Manchester said it could turn the feature on at a later date.
After 2023, universities moved in both directions. Jisc wrote in September 2023 that some institutions that had enabled the detector were disabling it, and others were doing the reverse. Its update of June 2025 said the general mood had shifted decisively away from automated AI detection.
Humanize your own paper
Transform your AI-assisted text and make it sound human, without touching important words or citations.
What does the Russell Group AI policy say?
The Russell Group's position is a set of five principles on generative AI in education, published on 3 July 2023. Member universities write their own policies and guidance, and the principles guide them. The five principles say:
- Universities will support students and staff to become AI-literate.
- Staff should be equipped to support students to use generative AI tools effectively and appropriately in their learning experience.
- Universities will adapt teaching and assessment to incorporate the ethical use of generative AI and support equal access.
- Universities will ensure academic rigour and integrity is upheld.
- Universities will work collaboratively to share best practice as the technology and its application in education evolves.
The principles don't mention AI detection tools. Instead, they cover issues such as AI literacy, assessment that allows ethical use, fair access (tools behind paywalls), and integrity. Integrity calls for universities to encourage students to be able to question and openly discuss their own use of AI tools without being penalized.
The principles say appropriate use is likely to differ between disciplines, and universities will encourage departments to apply the institution-wide policy in their own context. That is why two courses at the same Russell Group university can set different AI rules. Our university AI policy comparison shows how Oxford, Cambridge and other universities hand the decision to departments and assessment briefs.
How UK university AI rules treat coursework
The University of Glasgow's guidance for 2026-27 is a clear example of the rules a UK student now meets. It sorts every assessment into one of two scenarios, and the assessment brief should say which one applies.
| Glasgow scenario | Examples | Generative AI use |
|---|---|---|
| Supervised assessment | Exams, in-class tests, practical assessments, oral presentations, OSCEs | Not permitted unless the brief expressly allows it; AI may be used to prepare or revise |
| Unsupervised assessment | Essays, reports, portfolios, projects, dissertations | Generally permitted with acknowledgement; the course coordinator can limit or ban it for a given assessment |
Where a Glasgow brief says nothing, the default for coursework is that AI use is permitted with acknowledgement. Submitting AI-generated content as your own without acknowledgement, or when the AI contribution is substantial, is academic misconduct in both scenarios.
The acknowledgement Glasgow asks for has four parts. You state that you used generative AI, which tools you used (name, version and publisher), how you used them and what original work you contributed. Wording for that statement is in our AI disclosure statement templates.
Jisc's 2025 update on AI detection and assessment notes the rise of AI assessment scales and traffic-light systems, which tell students how much AI a task allows. Whatever system your course uses, the assessment brief is the document to read first.
What happens if a UK university suspects AI use
A suspicion of AI use goes through the university's academic misconduct procedure. The Office of the Independent Adjudicator for Higher Education (OIA), the independent student complaints body for England and Wales, describes what a fair process includes. Its casework note on AI and academic misconduct, published in July 2025, says almost all the AI complaints it receives come from students who went through a misconduct procedure.
- You should be told in writing what you may have done and why the university thinks so.
- You should get notice of meetings and all the relevant evidence, including detection software reports.
- The university has to prove the case, and the OIA says it is not on you to disprove it.
- Decision-makers should understand the strengths and limitations of detection software and weigh it against other information.
- You should get the chance to explain how you worked, and supplying notes and drafts can help.
- The university should consider whether assumptions about AI use could be biased against your writing style. The OIA gives disability, a communication difference and English as a second language as examples.
- Findings are made on the balance of probability, with clear reasons and an explanation of the penalty chosen.
Universities may also hold a viva, an oral discussion of the work you submitted. The OIA says the focus should be the content of the assessment, and the time since you completed the work should usually be taken into account. It calls an educational approach to minor or first instances good practice.
The OIA names English as a second language because detectors have been shown to flag non-native English writing more often. The evidence is in the post on why AI detectors are biased against non-native English writers.
How to appeal an AI finding at a UK university
An appeal starts inside the university. You use its academic appeals or complaints procedure within its time limits, and your students' union or students' association can help you prepare. An independent body can look at your case only after that procedure ends.
In England and Wales, the next step is a complaint to the OIA. Its casework note works as a checklist. You should get written notice of the allegation, all the evidence including any detector report, a fair chance to respond and clear reasons for the decision.
Once you've gone through the university's complaints or academic appeals procedure, in Scotland your complaint about a university or college is handled by the Scottish Public Services Ombudsman (SPSO). They can check if the procedures were reasonable and followed properly, including the process for academic appeals and student discipline. The SPSO can't review academic judgment or alter grades. They normally won't investigate issues from more than a year before you first learned about them.
The evidence to include with the letter is listed in the AI detection appeal letter template. The types of evidence that the OIA suggests that universities consider are draft history, version files, notes and sources. These show how your text has developed. The guide for the first meeting includes what to say and what to bring if you were accused of using AI but didn't use it.
Where TextPulse fits
TextPulse doesn't run an AI detector and doesn't sell detector scans, so it can't tell you what Turnitin would show your university. After every humanization it shows a Human Score from 0 to 100. That score is an estimate made in the app, and no detector is used to compute it.
When your course allows AI assistance and you acknowledge it, the TextPulse AI Humanizer is designed to lower AI scores on your drafts. It needs at least 50 words per humanization and takes up to 500 words per pass on Pro and 1,000 on Plus. The Document Humanizer, on Plus, takes a whole English .docx of up to 15,000 words and preserves its headings, lists, tables, captions and references.
Whatever AI detection UK universities use, keep your drafts and version history. A UK misconduct procedure looks at how you worked, and your drafts are the record of that.
Frequently Asked Questions
Some do and some don't, because each university decides for itself. Imperial College London, the University of Glasgow, Newcastle University and the University of Westminster say on their own pages that they don't use AI detection. The University of Bristol uses Turnitin's AI writing reports and tells staff to treat a result as supporting evidence reviewed by subject experts.
PhD in natural language processing, with years spent building NLP applications end to end. Moe works on text analysis: lexical and syntactic structure, and what separates machine-generated prose from human prose statistically. He has been experimenting with computational linguistics since the early days of NLTK, spaCy and WordNet, and still writes most of his tooling in Python.