Universities Dropping AI Detection: A Sourced Policy Tracker

As of 26 September 2026, 130 of the 148 universities and university units in our dataset have disabled, never enabled, restricted or discouraged AI detection. That is 35 disabled, 50 never enabled, 9 restricted and 36 discouraged. The other 18 still use a detector with official caveats (16) or have switched one back on (2). The dataset covers 11 countries and territories and records documented positions. It is not a random sample, so the 130 should not be read as a share of all universities.

The headline cases:

Vanderbilt disabled Turnitin's AI detector on 16 August 2023 "for the foreseeable future" (Vanderbilt).
Waterloo announced in May 2025 that it would discontinue Turnitin's AI detection, effective September 2025 (University of Waterloo).
Curtin disabled Turnitin's AI writing detection across all campuses from 1 January 2026 (Curtin University).
Washington State University (WSU) cancelled its contract for Turnitin's AI detection software in February 2026 (WSU Provost).
NTU in Singapore told faculty in August 2026 that it will deactivate its institutional AI detector on 31 December 2026 (CNA).
Oxford: "The University does not currently endorse any digital AI detectors for use in academic decision-making," its AI Competency Centre wrote in February 2026 (University of Oxford).

Every row in the tracker below links to its source, and the full dataset is free to download and reuse.

How we built this

Sources. Every row cites an official university page (119 rows), a policy document (23, all PDFs except one Word file) or a news report that quotes a named university official (6). Published Freedom of Information answers from Imperial, UEA and Hull count as official sources. One row, Robert Gordon University, cites the university's own case study published by the Quality Assurance Agency.
Every quote was re-checked. Each row carries a verbatim quote from its source. Every quote was checked word for word against that source on 26 September 2026, and all but one were re-checked in a second pass. The exception is Monash, whose page blocked the re-check; that row is marked "re-check blocked". Where a live page could not be fetched at all, the wording was checked against an Internet Archive capture instead; those rows are marked "archive-checked". Three rows link to an archive copy because the original page has been removed (Montclair State, UC Irvine and the University of Missouri System); their current status should be re-confirmed.
Status definitions. Each row has one of six statuses:
- disabled: the detector was on, or on by default, and the institution switched it off. NTU is counted here because it has announced its switch-off date, 31 December 2026.
- not-enabled: never switched on, opted out before launch, or never licensed.
- restricted: a detector may be available, but policy limits how results are used, for example never as the sole evidence.
- discouraged: the institution provides or recommends no detector and advises staff not to use one.
- enabled-with-caveats: the detector is on, with official guidance that scores are indicative only.
- re-enabled: switched back on after being turned off.
Where a statement could fit two categories, the dataset's notes column says so.
Dates. The Date column gives the date a change took effect, where the source states one. A second date in brackets refers to a related announcement, report or document. "As of" gives the date of the statement itself when no effective date is stated, and "undated" means the source carries no date. 99 of the 148 rows have no stated effective date.
Each status is as of its source's date. Policies change. DePaul and NC State switched detection back on, Mount Royal's AI working group is recommending a switch-off, and several rows rest on statements from 2023. Undated pages were online when we checked them on 26 September 2026; the three archive copies are the exception.
Absence from the list is not evidence of a position. An institution may be missing because it has not published a statement or because we have not found one. Its absence tells you nothing about what it does.
Unit-level rows are labelled. Three rows describe one school or division rather than a whole university: Arizona State's College of Liberal Arts and Sciences, MIT Sloan and UC San Diego's Division of Extended Studies. Southern Cross University's row covers research-degree candidates only.
Disclosure. We make AI writing software, including an AI detector and a humanizer. The statuses report what each institution says, not our view of any tool.

Download the data

university-ai-detection-policies.csv (CSV)
university-ai-detection-policies.json (JSON)

Both files contain all 148 rows and 12 columns: institution, country, status, detector, effective_date, announced_date, summary, quote, source_url, source_type, verified_on and notes. The notes column records how each quote was verified and any caveats. Dates are written as year-month-day or year-month, or unknown.

The data is licensed under CC BY 4.0. You may share and adapt it for any purpose, including commercially, as long as you credit "SupWriter (supwriter.com)", link to the licence and indicate any changes. Last verified: 26 September 2026. To report an error, or a policy we have missed, send the source link through our contact page.

Why universities turned it off

Most institutions explain their decision in a sentence or two. Three set out their evidence in detail.

Vanderbilt did the arithmetic on false positives. Its announcement noted that Turnitin claimed a 1% false-positive rate at launch, then put that in context: "Vanderbilt submitted 75,000 papers to Turnitin in 2022. If this AI detection tool was available then, around 750 student papers could have been incorrectly labeled as having some of it written by AI" (Vanderbilt, 16 August 2023).

Waterloo tested the tool itself. Internal testing by its Instructional Technologies and Media Services group "found the advantages of the detection tool were inconclusive. For example, in more than one instance the product flagged human written text as 100% generated by AI, which is concerning." Waterloo concluded that "the costs associated with Turnitin's AI detection feature outweigh the benefits" (University of Waterloo).

Washington State published outcomes. "Between 2023-2025, 33% of all Review Board cases related to allegations of inappropriate AI use led to a finding of not responsible because AI detection was submitted independent of any other supporting evidence." WSU also estimated that even at a 1% false-positive rate, "almost 1,485 assessments" of the 148,547 it ran through Turnitin in Fall 2024 "were likely flagged by Turnitin as AI generated when they were not" (WSU Provost).

Other reasons recur across the tracker:

Bias. Vanderbilt and Waterloo both cite evidence that detectors are more likely to flag writing by non-native English speakers. The best-known study found that seven detectors flagged, on average, 61% of 91 TOEFL essays by non-native writers as AI-generated (Liang et al., 2023).
Cost. At San Francisco State, Turnitin said the feature would become an add-on "for additional cost" when the free trial ended on 1 June 2024, and the score disappeared from reports (SF State). Western University dropped the tool in January 2024 "after it became a paid feature" (Western Gazette).
Privacy. Dalhousie says Nova Scotia privacy law and its own policy "do not allow you to submit or upload students' work to any tool, whether driven by AI or not, for AI detection" (Dalhousie Libraries). UVic says such tools are unreliable and that using them "violates the university's privacy policy" (UVic).
Turnitin's own caveat. Turnitin says the percentage "should not be used as the sole basis for action or a definitive grading measure by instructors" (Turnitin).

Complaint bodies and courts have reached similar conclusions. In a July 2025 case summary, the Office of the Independent Adjudicator, which handles student complaints about higher education providers in England and Wales, partly upheld a complaint from an international student called to a viva over a high Turnitin AI score. It found the provider "did not consider whether Turnitin's AI detection might be less reliable for non-native English speakers" (OIA). In January 2026 a New York court annulled Adelphi University's finding against a student whose essay Turnitin had scored as 100% AI-generated. The court called the finding "without valid basis and devoid of reason" and noted that the student's evidence, results from two other AI detectors rating the essay human-written, had not even been considered (decision on FindLaw).

Some universities went the other way. DePaul turned AI detection off for all its Turnitin accounts on 28 September 2023. "The decision was made to turn the tool back on in January 2024, prior to the beginning of Winter Quarter" (DePaul). NC State announced on 2 September 2025 that "Turnitin AI Detection is re-enabled", adding that "results can include false positives" (NC State DELTA). Sixteen institutions in the tracker keep a detector switched on with official caveats.

The full tracker

All 148 rows, grouped by country. Status and Date follow the definitions above; the full quote and notes for every row are in the downloadable files. In the Source column, "PDF" marks a document; "Internet Archive copy" links to an archived page because the original was removed; "archive-checked" means the live page blocks automated access, so the wording was checked against an archive capture; and "re-check blocked" marks the one row that could not be re-checked.

United States

42 entries: 19 disabled, 10 not-enabled, 7 discouraged, 4 enabled-with-caveats, 2 re-enabled.

Institution	Status	Date	What changed	Source
American University	disabled	2023 (reported 2023-12-06)	Disabled Turnitin's AI detection scores and will not base integrity investigations on any AI-detector score.	The Eagle (student newspaper)
Arizona State University (The College of Liberal Arts and Sciences) (unit-level)	discouraged	undated	College guidance: ASU recommends no AI-detection tools, and suspected AI use alone cannot start a formal case.	asu.edu
Baylor University	disabled	2023-12-31 (announced 2023-12-11)	Let Turnitin's AI detection feature expire at the end of 2023, citing too much uncertainty about accuracy.	baylor.edu
Boston University	disabled	2023-07-03	Disabled the Turnitin AI score panel, citing false positives and a lack of evidence for its efficacy.	bu.edu
Colorado State University	not-enabled	2023-04 (announced 2023-04-26)	Joined Turnitin's suppression list at launch, pausing AI writing detection over transparency and timing concerns.	colostate.edu
Cornell University	discouraged	undated	Teaching centre does not recommend automatic AI detectors for integrity cases, citing unreliability and no definitive evidence.	cornell.edu
DePaul University	re-enabled	2024-01	Turned AI detection off for all Turnitin accounts on 2023-09-28, then back on in January 2024.	depaul.edu
Georgetown University	disabled	2023-10	Turned off Turnitin's AI writing detection after the faculty executive committee and Honor Council raised accuracy concerns.	georgetown.edu
Indiana University	not-enabled	as of 2026-01-12	IU's Turnitin setup does not include the AI detection tool; guidance says not to upload work to detectors.	indiana.edu
Johns Hopkins University	disabled	as of 2025-03-06	Teaching guidance says JHU disabled Turnitin's AI detection over false positives; JHU has no official detection policy.	jhu.edu (archive-checked)
Michigan State University	not-enabled	undated	D2L help says Turnitin AI detection is not available at MSU and detectors should not be the sole basis.	msu.edu
MIT Sloan School of Management (unit-level)	discouraged	undated	School guidance says AI detectors have high error rates and recommends alternatives to detection.	mit.edu
Montclair State University	disabled	2023-11-20 (announced 2023-11-14)	Discontinued Turnitin's AI detection after a Provost memo; the original page has since been removed.	Internet Archive copy of montclair.edu
New York University	disabled	undated	Library guide states that AI writing detection has been disabled in Turnitin Similarity Reports.	nyu.edu
North Carolina State University	re-enabled	2025-09-02	Re-enabled Turnitin AI detection, telling instructors results can include false positives and should start conversations.	ncsu.edu
Saint Joseph's University	disabled	2023-09 (reported 2023-10-04)	Removed Turnitin's AI detection over false positives; its systems manager put its accuracy at 50 to 75%.	The Hawk (student newspaper)
San Francisco State University	disabled	2024-06-01 (announced 2024-05-01)	Stopped showing Turnitin's AI detection score when the free trial ended and it became a paid add-on.	sfsu.edu
Southern Methodist University	disabled	2023-12-31 (announced 2023-12-13)	Let its temporary Turnitin AI detection licence expire, citing limited success and a negative track record.	smu.edu
Syracuse University	disabled	as of 2025-06-24	Used a free preview in 2023; ITS now licenses no AI detector, and detector-only cases are not accepted.	Syracuse ITS (atlassian.net)
University of Alabama	disabled	2023-08-01	Removed Turnitin's AI writing detection from Blackboard after a temporary trial, based on feedback.	ua.edu
University of Arizona	disabled	2023-06-27	Temporarily disabled Turnitin's AI detection for all courses after inconsistent results; no reversal found on the page.	arizona.edu
University of California, Berkeley	not-enabled	2025-01 (first notice 2023-05-01)	Opted out at launch, ran an opt-in pilot from Fall 2023, then opted out of the tool entirely.	berkeley.edu
University of California, Irvine	not-enabled	2023	Integrity advisory committee decided Turnitin's AI detection would not be made available at UCI.	Internet Archive copy of uci.edu (PDF)
University of California, Los Angeles	not-enabled	2023-04	Opted out of Turnitin's AI detection preview in 2023; a 2025 UCLA post says it was never adopted.	ucla.edu
University of California, San Diego (Division of Extended Studies) (unit-level)	disabled	2025-04-07	Extended Studies deactivated Turnitin's AI detection; plagiarism detection remains available to instructors.	ucsd.edu
University of Colorado Boulder	not-enabled	undated	Chose not to add Turnitin's AI detection to its subscription after weighing effectiveness, cost and feedback.	colorado.edu
University of Iowa	discouraged	2024-09-30	Teaching office tells instructors to refrain from using AI detectors because of their inaccuracy.	uiowa.edu
University of Kansas	enabled-with-caveats	undated	Turnitin's AI detector is available in Canvas; teaching centre says it provides information, not an indictment.	ku.edu
University of Kentucky	enabled-with-caveats	2023-04	Provost's office said the AI score could not be switched off and urged healthy skepticism (2023 guidance).	uky.edu
University of Maryland, College Park	not-enabled	undated	IT knowledge base states that Turnitin AI detection is not enabled at UMD.	umd.edu
University of Michigan	discouraged	undated	Central faculty guidance does not recommend AI-detection technology, citing its high error rate.	umich.edu
University of Michigan-Dearborn	not-enabled	2023-04	Requested and received an opt-out from Turnitin's AI detection add-on when it launched.	umdearborn.edu
University of Missouri System	not-enabled	2023	All four campuses declined Turnitin's AI detection preview; guidance strongly discouraged detectors. Page since removed.	Internet Archive copy of missouri.edu
University of Oregon	discouraged	undated	Teaching office does not support AI detection tools; Student Conduct says detector percentages are not dispositive alone.	uoregon.edu
University of Pittsburgh	disabled	2023-06-22	Teaching Center disabled Turnitin's AI detector and does not endorse any AI-detection tool.	pitt.edu
University of Texas at Austin	discouraged	undated	Does not endorse AI detection software, holds no central contract, and bars third-party detectors without a university contract.	utexas.edu
University of Washington	disabled	undated	Asked Turnitin to switch off AI writing detection in UW's SimCheck integration, citing false positives and bias.	uw.edu
University of Wisconsin-Whitewater	enabled-with-caveats	2026-01-15	Turnitin's AI indicator available; scores under 20% are unreliable and never the sole evidence of misconduct.	uww.edu
Vanderbilt University	disabled	2023-08-16	Disabled Turnitin's AI detector for the foreseeable future, citing false positives and bias against non-native writers.	vanderbilt.edu
Washington State University	disabled	2026-02 (announced 2026-02-11)	Cancelled its Turnitin AI-detection contract; in 33% of AI cases, detector-only evidence led to not-responsible findings.	wsu.edu
West Chester University of Pennsylvania	enabled-with-caveats	undated	AI writing detection is enabled; guidance urges caution and warns against using it as the sole basis.	wcupa.edu (PDF)
Yale University	disabled	undated	Poorvu Center says Turnitin's AI detection feature is disabled and it does not support AI detection tools.	yale.edu

United Kingdom

43 entries: 3 disabled, 27 not-enabled, 1 restricted, 7 discouraged, 5 enabled-with-caveats.

Institution	Status	Date	What changed	Source
Abertay University	enabled-with-caveats	undated	Uses Turnitin's AI Writing Indicator; results must not be the sole evidential basis for a misconduct decision.	abertay.ac.uk
Aberystwyth University	disabled	2023-09-30 (announced 2023-09-25)	Turned off Turnitin's AI detection after an Academic Board discussion, citing false positives and student anxiety.	aber.ac.uk
Bangor University	discouraged	2026-09-01	Guidance says staff should not use AI detection when marking; not normally used in integrity cases.	bangor.ac.uk (Word document)
Bath Spa University	not-enabled	undated	Will not use digital AI-detection tools, including Turnitin's, because they have not proved reliable enough.	bathspa.ac.uk
Birkbeck, University of London	enabled-with-caveats	as of 2026-04	Staff use Turnitin's AI checker, then check any flag against permitted AI use and the student's declaration.	bbk.ac.uk (PDF)
Bournemouth University	not-enabled	2023-04 (announced 2023-04-13)	Opted out of Turnitin's AI detection launch, saying it was unfair without testing, guidance or policy.	bournemouth.ac.uk
Brunel University of London	not-enabled	undated	Not planning to use Turnitin's AI detection tool, citing easy evasion and unacceptable false-positive rates.	brunel.ac.uk
Durham University	not-enabled	undated	Chose not to enable Turnitin's AI detector, saying it does not fit its approach to assessment.	durham.ac.uk
Edinburgh Napier University	not-enabled	as of 2025-08	Leadership does not permit AI detection on submissions, citing unreliability, bias and lack of student consent.	napier.ac.uk (PDF)
Goldsmiths, University of London	not-enabled	undated	Library guide says Goldsmiths has not enabled Turnitin's AI detection, mainly over accuracy and fairness concerns.	gold.ac.uk
Heriot-Watt University	not-enabled	AY 2023-24 only	Said it would use no generative-AI detection tools in 2023-24; no later statement was found.	hw.ac.uk (PDF)
Imperial College London	not-enabled	as of 2026-09	FOI page: does not use AI detection software such as Turnitin's and has never trialled or enabled any.	imperial.ac.uk (FOI page)
King's College London	not-enabled	as of 2023-10	Decided not to enable Turnitin's AI detection percentage over reliability and false-positive concerns.	kcl.ac.uk
Manchester Metropolitan University	not-enabled	undated	Opted out of Turnitin's AI detector for the foreseeable future; staff must not upload work to online detectors.	mmu.ac.uk
Middlesex University	discouraged	2026-09	New AI policy says detectors give too many false positives to judge misconduct; favours assessment design.	mdx.ac.uk (PDF)
Newcastle University	disabled	undated	Does not support automatic AI checkers; the AI detection features of its platforms have been disabled.	ncl.ac.uk
Northumbria University	enabled-with-caveats	as of 2024-03	2024 FAQ: Turnitin's AI detector gives tutors data but does not decide whether misconduct occurred.	northumbria.ac.uk (PDF)
Queen's University Belfast	discouraged	as of 2025-10	2025-26 guidance says text-based AI detectors give unacceptably high false positives and are not recommended.	qub.ac.uk (PDF)
Robert Gordon University	enabled-with-caveats	2023-10 (case study 2025-10)	Switched on Turnitin's AI detection despite reliability concerns, as one input alongside other evidence.	qaa.ac.uk (RGU case study, PDF)
University College London	not-enabled	as of 2024-10-08	Student guidance says UCL does not use GenAI detectors when marking; markers ask for a chat instead.	ucl.ac.uk
University of Bath	not-enabled	2026-05-11 (guide 2026-04-16)	Uses Inspera Originality, not Turnitin; AI authorship detection is not enabled in its implementation.	bath.ac.uk
University of Birmingham	not-enabled	as of 2023-07	Has not enabled Turnitin's AI writing detection; using any AI detection tool is currently not allowable.	birmingham.ac.uk
University of Bristol	enabled-with-caveats	as of 2025-11-20	Turnitin AI writing reports are on; subject experts must review results, which support rather than replace academic judgement.	bristol.ac.uk
University of Cambridge	discouraged	undated	Does not encourage AI detection software, citing proven inaccuracy, and says detectors provide no evidence.	cam.ac.uk
University of Dundee	not-enabled	2023-04	Opted out of Turnitin's AI detection launch because it could not test reliability; bars unauthorised detectors.	dundee.ac.uk
University of East Anglia	not-enabled	as of 2026-09-24	FOI reply: no AI detection tool in use or ever trialled; a 2023 UEA policy approved Turnitin's for limited use.	uea.ac.uk (FOI reply)
University of Edinburgh	not-enabled	as of 2025-03-28	Project report says Turnitin's AI detection, added in April 2023, is currently switched off.	ed.ac.uk
University of Glasgow	not-enabled	2023-04	Took up Turnitin's AI detection opt-out along with most Russell Group universities, at least for the moment.	gla.ac.uk
University of Greenwich	not-enabled	2023-03-31	Opted out of Turnitin's AI detection because its effect on student outcomes was unknown.	gre.ac.uk
University of Hull	not-enabled	undated	FOI reply to a July 2025 request: holds no information because it does not use AI detection tools.	hull.ac.uk (FOI reply, PDF)
University of Leeds	discouraged	undated	Staff guidance: do not use online AI detection tools, which are too inaccurate to be used as evidence.	leeds.ac.uk
University of Liverpool	disabled	2024-08-01 (announced 2024-07-18)	Piloted Turnitin's AI detection from September 2023, then disabled it after finding lower-than-expected reliability.	liverpool.ac.uk
University of Manchester	restricted	2025-02-14	University AI guidelines say detector output is unreliable and biased and cannot be used as malpractice evidence.	manchester.ac.uk
University of Nottingham	not-enabled	2023-04 (confirmed 2024-04-15)	Opted out in April 2023; after testing found many false positives, declined to implement the tool.	nottingham.ac.uk (archive-checked)
University of Oxford	discouraged	as of 2026-02-16	AI Competency Centre: the University does not endorse any digital AI detectors for academic decision-making.	ox.ac.uk
University of Portsmouth	not-enabled	undated	Staff AI guidance summary states that AI detection tools are not currently allowed.	port.ac.uk
University of Reading	not-enabled	as of 2023-11	Guidance says Turnitin's AI detection is turned off over low reliability and the risk of false accusations.	reading.ac.uk (PDF)
University of Sheffield	not-enabled	undated	Study-skills guidance says GenAI detection tools are not used, citing error rates and false positives.	sheffield.ac.uk
University of South Wales	not-enabled	as of 2024-03	Formal statement: chose not to implement any AI detection tools, citing false positives, opacity and bias.	University of South Wales (PDF)
University of St Andrews	not-enabled	as of 2025-01-28	Student FAQs say Turnitin's AI detection has not been switched on, over concerns about accurate detection.	st-andrews.ac.uk (PDF)
University of Strathclyde	not-enabled	2025-08	Senate-approved guidance does not permit Turnitin's AI Checker or any other AI detector on student work.	strath.ac.uk (PDF)
University of Sussex	discouraged	undated	Staff told not to upload work to online AI detectors; their outputs are not admissible as misconduct evidence.	sussex.ac.uk
University of the Arts London	not-enabled	undated	Does not use AI detection software, saying it is not accurate enough to be reliable evidence.	arts.ac.uk (archive-checked)

Canada

26 entries: 2 disabled, 5 not-enabled, 1 restricted, 18 discouraged.

Institution	Status	Date	What changed	Source
Brock University	discouraged	2024-05 (memo 2025-01-28)	Has not procured AI detection software; instructors should not use detectors, and their reports cannot be evidence.	brocku.ca
Carleton University	restricted	as of 2026-08	Briefing note: AI detection tools are unreliable and are not accepted as sole evidence of violations.	carleton.ca (PDF)
Concordia University	not-enabled	undated	Turnitin's AI Writing tab is not included in Concordia's licence, so it is unavailable in reports.	concordia.ca
Dalhousie University	discouraged	undated	Library guide: provincial privacy law and university policy bar uploading student work to AI detection tools.	Dalhousie Libraries (libguides.com)
McGill University	discouraged	undated	Teaching guidance says McGill discourages AI detection tools, citing unreliability, false positives and classroom stress.	mcgill.ca
McMaster University	discouraged	as of 2024-08	Provost guidelines call automatic AI detectors unreliable and not recommended; McMaster declined Turnitin's detector in 2023.	mcmaster.ca
Memorial University of Newfoundland	discouraged	as of 2025-06-02	Senate-level statement: GenAI detection tools are discouraged and cannot be the only evidence for an allegation.	mun.ca (PDF)
Mount Royal University	discouraged	as of 2026-08	Working group advises avoiding AI detection and is recommending that Turnitin's AI tool be disabled.	mtroyal.ca
Queen's University	not-enabled	2023-04 (page 2023-07)	Turnitin AI reports withheld from instructors while Queen's evaluates the tool; the 2023 page is still live.	queensu.ca
Simon Fraser University	not-enabled	undated	Turnitin's AI detector is not enabled because its accuracy and reliability have not been independently evaluated.	sfu.ca
Toronto Metropolitan University	discouraged	undated	Academic Integrity Office endorses no AI detection tools, calling them unreliable and a possible privacy concern.	torontomu.ca
University of Alberta	discouraged	undated	Provost taskforce guidance generally does not recommend AI detection apps; any exception needs a privacy review.	ualberta.ca
University of British Columbia	not-enabled	2023-04 (affirmed 2023-08-28)	Decided not to enable Turnitin's AI detection and not to centrally install any AI-detection tool.	ubc.ca
University of Calgary	discouraged	undated	Student Success Centre says UCalgary does not encourage AI detectors, calling them unreliable.	ucalgary.ca
University of Guelph	not-enabled	as of 2024-09-30	AVP Academic memo: GenAI detection tools not authorized, and Turnitin's AI feature has not been turned on.	University of Guelph (memo, PDF)
University of Manitoba	discouraged	undated	Teaching centre says the university discourages AI-detection software, citing unreliability and privacy concerns.	umanitoba.ca
University of Ottawa	discouraged	as of 2023-04-18	Teaching support service discourages AI detectors on student work because their quality is unconfirmed (2023 guidance).	uottawa.ca
University of Saskatchewan	discouraged	as of 2026-05	Integrity site says GenAI detection tools are unreliable and that none has been approved for use.	usask.ca
University of Toronto	discouraged	as of 2025-06-18	Does not support AI-detection software on student work, citing unreliability and false flags of human writing.	utoronto.ca
University of Victoria	discouraged	undated	Policy FAQ: instructors may not use AI tools to detect violations, and results are not accepted as evidence.	uvic.ca
University of Waterloo	disabled	2025-09 (announced 2025-05)	Discontinued Turnitin's AI detection, citing unreliability, bias against non-native writers, internal testing and cost.	uwaterloo.ca
University of Windsor	discouraged	undated	AI guidance says detectors are not permitted, citing low efficacy, bias and the risk of false accusations.	uwindsor.ca
Université Laval	discouraged	undated	Teaching support service advises against AI-detection software, citing reliability limits and ethical concerns.	ulaval.ca
Western University	disabled	2024-01 (reported 2024-01-18)	Used Turnitin's AI detection in a free preview, then stopped once it became a paid feature.	Western Gazette (student newspaper)
Wilfrid Laurier University	discouraged	undated	Endorses no generative-AI detection tools; their outputs are not accepted as evidence of misconduct.	wlu.ca
York University	discouraged	undated	Does not currently recommend AI detection tools and tells instructors not to rely on them as evidence.	yorku.ca

Australia

20 entries: 8 disabled, 4 not-enabled, 4 restricted, 1 discouraged, 3 enabled-with-caveats.

Institution	Status	Date	What changed	Source
Australian Catholic University	disabled	2025-03-17 (announced 2025-03-12)	Piloted switching off Turnitin's AI indicator over reliability concerns; ABC later reported ACU abandoned the tool.	acu.edu.au (archive-checked)
Australian National University	disabled	2024-01-01	Made Turnitin's AI writing detection available from April 2023, then disabled it for all staff.	anu.edu.au
Charles Sturt University	disabled	2024-01-01 (announced 2023-10-19)	Phased out AI text detection, ending access to Turnitin's tool and any other AI text detector.	csu.edu.au (archive-checked)
Curtin University	disabled	2026-01-01 (announced 2025-09-04)	Turnitin's AI writing detection disabled across all campuses and study periods; text-matching checks remain.	curtin.edu.au
Deakin University	not-enabled	undated	Declined to apply Turnitin's AI detector at its 2023 launch; says it does not use AI-detection software.	deakin.edu.au
Federation University Australia	restricted	as of 2026-04-01	Integrity procedure: AI detector results alone are not evidence of misconduct, and allegations need corroboration.	federation.edu.au
James Cook University	not-enabled	as of 2026-08-13	DVC Education says JCU chose not to invest in AI detection tools because they are unreliable.	jcu.edu.au
Macquarie University	disabled	Session 2, 2023	Disabled Turnitin's AI detection after it had been switched on by default at the April 2023 launch.	mq.edu.au
Monash University	discouraged	undated	No AI detection tools approved; detector results cannot be used as evidence when reporting a suspected breach.	monash.edu (re-check blocked)
Murdoch University	disabled	as of 2026-05-28	Disabled AI detection reports in Turnitin to align with other Western Australian universities; originality checks continue.	murdoch.edu.au
Southern Cross University (research degrees only)	restricted	as of 2026-09-10	Research-degree guidelines: supervisors must not rely on AI-detector output as the sole basis for an allegation.	scu.edu.au (PDF)
University of Canberra	enabled-with-caveats	as of 2025-10-23	Has enabled Turnitin's AI writing detection; scores may prompt a conversation or be one piece of evidence.	canberra.edu.au
University of Melbourne	restricted	as of 2026-02-26	Turnitin's AI detection is available to staff, but a report alone is not sufficient evidence for an allegation.	unimelb.edu.au
University of Queensland	disabled	mid-2025	Disabled the Turnitin AI detection tool previously available to staff; calls AI detectors flawed and unreliable.	uq.edu.au
University of Sydney	restricted	as of 2026-06-02	Turnitin AI detection may be used when misuse is suspected, but the score is never the only evidence.	sydney.edu.au
University of Western Australia	not-enabled	as of 2026-02-16	DVC Education says UWA decided not to use any AI detection tools, calling them unreliable and inequitable.	uwa.edu.au
University of Wollongong	not-enabled	as of 2023-08-30	Education Committee minutes: UOW will not use AI detection tools while false positives remain too high.	uow.edu.au (PDF)
UNSW Sydney	enabled-with-caveats	2023-04-05 (guidance 2025-11-12)	Authorises only Turnitin's AI detector; a flag merely triggers investigation, and markers rely on their own judgement.	unsw.edu.au
Victoria University	disabled	2026 (Semester 1, Block 1)	Turnitin's AI Writing Detection feature switched off by default from the start of 2026 Semester 1, Block 1.	vu.edu.au
Western Sydney University	enabled-with-caveats	as of 2024-11	Staff use AI writing detection reports only as a prompt to look for other indicators of misuse.	westernsydney.edu.au

Ireland

5 entries: 1 disabled, 2 not-enabled, 1 discouraged, 1 enabled-with-caveats.

Institution	Status	Date	What changed	Source
Dublin City University	discouraged	as of 2025-03-06	Position statement tells staff not to use AI detection software on student work, citing poor effectiveness and data harvesting.	dcu.ie
Maynooth University	enabled-with-caveats	as of 2025-04-02	Turnitin gives staff a GenAI score that is not definitive proof but may start an integrity process.	maynoothuniversity.ie (PDF)
University College Cork	not-enabled	as of 2026-05-15	Integrity policy: AI detection software is not sanctioned for detecting or investigating alleged misconduct.	ucc.ie
University of Galway	not-enabled	undated	Does not authorise or advocate AI detectors because they have not been shown to be effective.	universityofgalway.ie
University of Limerick	disabled	2024-03 (announced 2024-03-07)	Decided not to sign up to Turnitin's AI add-in, in line with national (NAIN) guidance on detectors' limits.	ul.ie

New Zealand

3 entries: 1 not-enabled, 1 restricted, 1 enabled-with-caveats.

Institution	Status	Date	What changed	Source
Lincoln University (NZ)	restricted	as of 2025-12-02	Staff guidelines: Turnitin's AI detection is a screening tool only, and results must be verified before referral.	lincoln.ac.nz (PDF)
University of Auckland	not-enabled	2023-04-05	Piloted Turnitin's AI detection, then asked Turnitin not to enable it for its account at release.	auckland.ac.nz
University of Canterbury	enabled-with-caveats	as of 2026-09-24	Turnitin AI detection used in some courses as a first screening step; generally not the sole evidence.	canterbury.ac.nz

Singapore

4 entries: 1 disabled, 1 not-enabled, 2 restricted.

Institution	Status	Date	What changed	Source
Nanyang Technological University	disabled	2026-12-31 (scheduled; announced 2026-08-13)	Will deactivate its institutional AI detector at semester end, calling automated detection tools fundamentally unreliable.	CNA (news)
National University of Singapore	restricted	as of 2026-09-15	AI policy: detector verdicts are not admissible as evidence of dishonesty; iThenticate's AI detection is not enabled.	nus.edu.sg (PDF)
Singapore Management University	restricted	as of 2026-08-24	Still uses AI detection tools, but never as the sole basis for a misconduct finding.	CNA (news)
Singapore University of Technology and Design	not-enabled	as of 2026-08-24	Confirmed to CNA that it does not use automated AI detection tools or a campus-wide detector.	CNA (news)

Hong Kong

2 entries: 2 enabled-with-caveats.

Institution	Status	Date	What changed	Source
Hong Kong University of Science and Technology	enabled-with-caveats	as of 2023-12-27	Teaching centre calls Turnitin's AI detection unreliable and advises against using it as conclusive evidence or for punishment.	hkust.edu.hk
University of Hong Kong	enabled-with-caveats	undated	Says Turnitin's AI check produces false positives; teachers should not rely on it, and results are for reference only.	hku.hk

Other countries

3 entries: 1 disabled, 2 discouraged.

Institution	Status	Date	What changed	Source
Delft University of Technology (TU Delft), Netherlands	discouraged	undated	Teaching support says relying on AI detectors to penalise students risks false accusations and unfair treatment.	tudelft.nl
University of Cape Town, South Africa	disabled	2025-10-01 (announced 2025-07-24)	Stopped using AI detection tools such as Turnitin's AI Score under its AI in Education Framework.	uct.ac.za
EPFL (École polytechnique fédérale de Lausanne), Switzerland	discouraged	as of 2024-11	Teaching guide: no AI detectors are validated, and teachers should not rely on AI-detection results.	epfl.ch

What this means if you're a student

A detector being switched off changes how work is checked, not what you are allowed to do. Your course's rules on AI use still apply, and staff can still ask to talk through your work with you.

Check the current policy, not a list. Positions change, so confirm your own university's page and its date, even if it appears in this tracker.
Keep your process. Version history, notes and drafts show how your work developed. If you have been flagged, how to prove your writing is human walks through the evidence and the appeal steps.
Understand false positives. Why human writing gets flagged as AI covers the research, including who is most at risk.
Don't assume a safe score. What percentage of AI is acceptable on Turnitin? looks at the thresholds institutions actually use.
Pre-checking a draft? Free Turnitin AI detector alternatives compares free checkers and their limits. None of them is Turnitin, so treat any score as a rough signal.

Frequently Asked Questions

Which universities have turned off Turnitin's AI detector?

In this tracker, 34 institutions have switched off AI detection after having it on (33 of them explicitly Turnitin's), and NTU will switch off its institutional detector on 31 December 2026. They include Vanderbilt (16 August 2023), the University of Liverpool (1 August 2024), Waterloo (September 2025), Curtin (1 January 2026) and Washington State (February 2026). Another 50 institutions never enabled a detector, among them the University of British Columbia and King's College London. The country tables list every institution with its date and source.

Why are universities turning off AI detection?

The reasons they give are mostly about accuracy and fairness. Vanderbilt estimated that a 1% false-positive rate could have mislabelled about 750 of the 75,000 papers it submitted to Turnitin in 2022. Washington State found that 33% of its AI-related Review Board cases in 2023-2025 ended in a finding of not responsible because detector results were submitted without other supporting evidence. Institutions also cite bias against non-native English writers, weak results in their own testing (Waterloo; Liverpool's pilot found "a lower-than-expected level of reliability"), cost once free access ended, and privacy rules on uploading student work. Turnitin itself says its percentage should not be the sole basis for action.

Has any university switched AI detection back on?

Yes. DePaul turned Turnitin's AI detection off on 28 September 2023 and back on in January 2024, and NC State announced on 2 September 2025 that it had re-enabled it. Others use it: 16 institutions in the tracker run a detector with official caveats, and Robert Gordon University switched Turnitin's on in October 2023 despite reliability concerns. Pilots have also ended in opt-outs: UC Berkeley offered Turnitin's AI detection through an opt-in pilot from Fall 2023 to the start of Spring 2025, then opted out.

How can I check my own university's policy?

Search your university's website for "Turnitin AI" or "AI detection" together with "academic integrity". The answer is usually in the academic integrity policy, the teaching centre's guidance or the help pages for your learning platform, which is where most sources in this tracker come from. Then read your course outline or assessment brief, because course rules on AI use can be stricter than university guidance. If nothing is published, ask your academic integrity office or your students' union advice service. Some universities also publish answers to Freedom of Information requests: Imperial, UEA and Hull have all answered questions about AI detection that way. Whatever you find, check its date. If your university is missing here or its entry is out of date, send us the link through our contact page.

Updates and corrections

This page was rebuilt on 26 September 2026 around the dataset above, replacing an earlier list of seven universities. Earlier versions gave wrong dates for Vanderbilt (correct: disabled 16 August 2023), Waterloo (announced May 2025, effective September 2025) and Curtin (announced 4 September 2025, effective 1 January 2026), and quoted statements we could not trace to the institutions; those quotations have been removed.

SupWriter makes AI writing tools, including an AI detector for pre-checking drafts. It is one scoring model, not the detector your university uses, and no tool can guarantee a result. 300 words free, no credit card required.