{"resourceId":"glasgow-guardian-turnitin-no-ai-training-student-work-20261017","versions":[{"version":"external-15e65f565cc7bd062f5ca0201c3985279de75c07dc45e2034e70bb57ac7e0a6e","resource":{"id":"glasgow-guardian-turnitin-no-ai-training-student-work-20261017","title":"Glasgow Guardian: University of Glasgow says Turnitin confirmed in writing no student submissions will train AI, after sector pushback over EULA change; Southampton exiting Turnitin","organization":"University of Glasgow; Turnitin; University of Southampton (as reported by the Glasgow Guardian student newspaper)","sector":"Higher education academic integrity tooling / student data protection / vendor contracting","geography":"United Kingdom (Glasgow, Scotland; Southampton, England)","publishedAt":"Glasgow Guardian, 17 October 2026 (reporting a 5 October university statement)","publicationDate":"2026-10-17","eventDate":"2026-10-05","sourceName":"Glasgow Guardian","sourceLabel":"Student newspaper report","sourceUrl":"https://glasgowguardian.co.uk/2026/10/17/university-of-glasgow-assures-students-no-submitted-work-will-be-used-to-train-ai/","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["data-security","governance-procurement"],"finding":"The paper reports that a Turnitin EULA update due 14 July would have let AI services use submitted student work to train models. In a 5 October statement, Glasgow said a coordinated UK sector response held that the change would breach current contract terms and end its use of the platform. Turnitin then confirmed in writing that no submitted work would train AI models. Glasgow says its data protection agreement means no Glasgow work can be used to train AI \"now or in the future\". Southampton will not renew after 2026/27 because Turnitin intends to revisit the change. Glasgow opted out of Turnitin AI detection in 2023 over false-positive concerns.","sledRelevance":"Integrity tools that hold large corpora of student work are a direct AI data-governance exposure. Contract terms and collective sector action, not policy statements alone, decided the outcome here.","evidence":"Student-newspaper reporting quoting university statements. The Turnitin written confirmation and contract terms are not published in the article.","architectureImplications":"Interpretation: plagiarism and detection platforms are data processors whose terms can change by EULA. Institutions need contract-level controls that override click-through terms.","governanceImplications":"Shows the value of sector-coordinated negotiation and of exit options (Southampton). Detector opt-out on false-positive grounds remains part of Glasgow's posture.","securityPrivacyImplications":"Student intellectual property and personal data in submissions could have been repurposed for model training. Retention and secondary-use terms are central.","caveats":"Secondary student reporting of UK events. Vendor position may be revisited next year per Southampton. Excludes U.S. contracts and FERPA analysis.","streamIds":["student-success"],"roles":{"sales":"Sales framing: integrity platforms hold years of student work, and a single EULA change nearly allowed AI training on it until UK universities pushed back collectively. This supports discovery on contract terms for integrity and LMS add-on tools and on who monitors term changes. Offer a bounded contract and data-flow review. Do not suggest that switching tools removes risk or that detection improves integrity outcomes; Glasgow itself opted out of AI detection over false positives.","engineering":"Engineering framing: treat plagiarism and AI-detection services as data processors with persistent repositories of student submissions. Pre-sales checks: where submissions are stored, repository opt-out options, secondary-use and model-training clauses, change-of-terms notice, deletion on exit, and export formats if an institution leaves, as Southampton plans. Detection features should be configurable off, consistent with institutions declining them over false-positive risk, and every setting change should be logged for audit.","delivery":"Delivery framing: set up a standing vendor-terms watch for integrity tools, an annual data-use attestation, and an exit plan that covers repository deletion and alternative workflows. Coordinate with sector consortia when negotiating. Communicate clearly to students what happens to their submissions. If detection is disabled, document the human-review integrity process that replaces it, with appeal routes, evidence standards and named decision-makers."},"roleDetails":{"sales":{"customerProblem":"Student work held by integrity vendors may be repurposed for AI training through unilateral licence changes.","stakeholders":["Academic integrity office","Data protection officer or privacy office","Procurement and legal","Student union or government","CIO"],"discoveryQuestions":["Do current integrity-tool contracts prohibit training on student submissions?","Who is notified when vendor EULAs change?","Is AI detection enabled, and on what evidence standard?","What would it take to exit the current platform?"],"valueHypothesis":"Interpretation: a contract and data-flow review can reduce exposure to secondary use of student work and give leverage for collective negotiation.","potentialEngagement":"A bounded integrity-vendor contract and data-flow review with exit-plan outline.","unsupportedClaims":["That any detector is reliable for misconduct findings","That switching vendors eliminates data risk","That Turnitin will not revisit its terms","U.S. legal conclusions"]},"engineering":{"fit":"Applies to any institution using cloud plagiarism or AI-detection services with stored repositories.","architectureIntegration":"LMS-integrated submission flow to a vendor repository. Configuration of repository storage and detection features.","prerequisites":["Inventory of integrity integrations","Copy of current contract and DPA","Repository settings review","Exit and export requirements"],"constraints":["UK GDPR context in source; U.S. institutions must map to FERPA and state law","Click-through EULAs may conflict with master contracts","Vendor may revisit terms in 2026/27"],"security":"Require no-training and no-secondary-use clauses, deletion on exit, access logs and subprocessor transparency.","proofOfValue":["Contract clause gap analysis","Repository data-retention verification","Test of export and deletion procedures","Configuration audit confirming detection settings match policy"]},"delivery":{"work":"Create a vendor-terms watch, contract amendments and an exit plan for integrity tooling.","dependencies":["Legal and procurement capacity","Vendor cooperation","Sector consortium coordination","Integrity policy owners"],"ownership":"Academic integrity office with the data protection officer and procurement.","skillsAdoption":"Staff need clarity on the evidentiary limits of similarity and detection reports. Students need plain-language notices about data use.","governanceCheckpoints":["Annual vendor data-use attestation","Change-of-terms review trigger","Detection-feature policy review","Exit-readiness review before renewal"],"acceptanceCriteria":["Proposed: written no-training clause in every integrity-tool contract","Proposed: 30-day notice requirement for material terms changes","Proposed: documented, tested deletion-on-exit procedure","Proposed: published student notice on submission data use"],"risks":["Unilateral vendor terms changes","Over-reliance on detection outputs","Disruption if exiting a platform mid-year","Student trust erosion"]}},"retrievedAt":"2026-10-19T03:26:10Z","enrichedAt":"2026-10-19T03:40:00Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Possible platform exits would require integrity-process and staff workflow changes. Not quantified.","procurementImplications":"Interpretation: add explicit no-training, no-secondary-use and change-of-terms notice clauses to integrity and LMS add-on contracts, with termination rights.","operatingModelImplications":"Integrity offices, procurement and data protection officers must jointly monitor vendor terms. Sector consortia can strengthen negotiating leverage.","sourceVerification":{"openedUrl":"https://glasgowguardian.co.uk/2026/10/17/university-of-glasgow-assures-students-no-submitted-work-will-be-used-to-train-ai/","referenceExcerpt":"Turnitin has since engaged with the sector and confirmed in writing that no submitted work will be used to train AI models.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}