{"resourceId":"wonkhe-studiosity-keeping-the-learning-real-integrity-20261004","versions":[{"version":"external-50adf533fe23c4fd944bec68c4d97ec99a3d5be6e1f26109a0c982f20ae02c06","resource":{"id":"wonkhe-studiosity-keeping-the-learning-real-integrity-20261004","title":"Wonkhe/Studiosity 'Keeping the learning real': UK sector moves from AI detection to integration, but validating learning in AI-permitted assessment remains the residual risk","organization":"Wonkhe with Studiosity (vendor co-author)","sector":"Higher education assessment and academic integrity / GenAI policy","geography":"United Kingdom (transferability to U.S. institutions limited)","publishedAt":"Report dated October 2026; launch post on Wonkhe dated 4 October 2026 (PDF last modified 2 October 2026)","publicationDate":"2026-10-04","eventDate":null,"sourceName":"Wonkhe","sourceLabel":"Sector 'temperature check' report co-produced with a vendor","sourceUrl":"https://wonkhe.com/media/2026/10/Keeping-the-learning-real.pdf","evidenceClass":"vendor-claim","outcomeClass":"mixed","topics":["governance-procurement","knowledge-work","accessibility-workforce","operating-model"],"finding":"Based on a desk review and three online round tables in September 2026 (two with senior institutional leaders and professional association representatives, one with student representatives), the report finds UK institutions shifting from bans and detection toward 'open' and 'closed' assessment mixes (traffic-light or Perkins-scale approaches), AI literacy, and validation of learning. It cites YouGov polling for Studiosity in which three-quarters of students reported stress over being wrongly accused of AI plagiarism. Students said institutional strategy has not always reached practice and that changes can seem arbitrary; vague briefs ('assist', 'brainstorm') burden students for whom English is not a first language and neurodivergent students. Closed assessments face wearable-device risks and inclusion concerns; vivas do not scale; declarations feel like policing when guidance is unclear. The authors place the residual risk in validating learning for open assessments. The report ends with a Studiosity product section.","sledRelevance":"Published this weekend, it is the sharpest current sector account of where AI assessment policy is stuck: permitting AI is easier than proving learning happened. It complements the Kent State employer experiment, which finds employers reassured mainly by in-person tests.","evidence":"Qualitative: desk review plus three round tables; no sample sizes, institution counts, or systematic coding reported. Quantitative claim (three-quarters stressed) is from separate YouGov polling commissioned by Studiosity and not detailed here.","architectureImplications":"Learning-validation steps (declarations, self-validation prompts, vivas) need workflow support in the LMS and assessment systems, including records of how validation was done.","governanceImplications":"Integrate AI literacy, open and closed assessment design, and validation as one framework; explain course-level variation to students and involve student representatives from the start.","securityPrivacyImplications":"Detection-led misconduct cases based on patchy evidence harm students; any validation or proctoring approach needs proportionate data collection and due-process safeguards.","caveats":"Co-produced with Studiosity, which promotes its own validation product in the report; qualitative round tables with unknown participant numbers; UK regulatory and assessment context (vivas, module structures) differs from U.S. practice; the YouGov statistic is vendor-commissioned.","streamIds":["student-success"],"roles":{"sales":"The framing point is that permitting AI is easier than proving learning happened. This UK sector review finds institutions moving away from detection, with validation of learning in AI-permitted assessments left as the open risk. Use it to start assessment-strategy conversations with provosts and integrity officers. Be explicit that it is vendor co-produced, qualitative and UK-based, and do not position any detection or validation product as a proven fix.","engineering":"For engineering, the framing point is that validation needs workflow, not detection. Support AI-use declarations, structured self-validation prompts and viva scheduling inside the LMS, and record which validation method applied to each assessment. Avoid detection tools as evidence for misconduct. Check accessibility for every validation path and keep data collection proportionate. Treat the vendor framework described in the report as one option to evaluate, not a reference architecture.","delivery":"Delivery's framing point is consistency students can understand. Work at programme level: map open and closed assessments, replace vague terms like 'assist' with concrete examples, and explain why each assessment takes its approach. Involve student representatives before drafting. Pilot a scalable validation step in one programme and track student anxiety, accommodation requests and staff time. Review the results with students before extending the approach to other programmes."},"roleDetails":{"sales":{"customerProblem":"Institutions have moved to permit AI in assessment but lack scalable, trusted ways to confirm that learning occurred, while students fear wrongful accusations.","stakeholders":["Provost / PVC education","Academic integrity office","Teaching and learning center","Student government or union","Disability services"],"discoveryQuestions":["How do programmes decide which assessments permit AI?","What evidence is used in AI misconduct cases?","Which validation methods (declarations, vivas) are in use, and at what staff cost?"],"valueHypothesis":"An integrated framework of literacy, assessment design and proportionate validation may reduce misconduct disputes and student anxiety; no measured outcomes here.","potentialEngagement":"Programme-level assessment review and validation-step pilot with student co-design.","unsupportedClaims":["That any product secures academic integrity","Reliable AI detection","U.S. transferability without adaptation"]},"engineering":{"fit":"Institutions with LMS-based assessment that permit AI in some tasks.","architectureIntegration":"LMS assessment settings → declaration and self-validation components → records of validation method linked to grades.","prerequisites":["Programme-level assessment map","LMS support for declarations and structured prompts","Accessibility review"],"constraints":["Vivas do not scale","Wearable devices undermine closed assessments","UK-derived practices"],"security":"Do not use detector scores as sole misconduct evidence; limit retention of declaration and validation data.","proofOfValue":["Student-reported clarity of AI guidance","Number of misconduct disputes","Staff time per validation step"]},"delivery":{"work":"Map assessments, rewrite briefs with concrete AI-use examples, co-design with students, and pilot a validation step in one programme.","dependencies":["Module leader time","Student representative participation","Disability services input"],"ownership":"Teaching and learning center with the academic integrity office.","skillsAdoption":"Staff model critical AI use; students build judgement through in-class discussion of AI expectations.","governanceCheckpoints":["Student co-design before policy drafting","Accessibility review of validation methods","Misconduct-evidence standards review"],"acceptanceCriteria":["Defined improvement in student-reported clarity of AI rules","Reduction in contested AI misconduct cases","Validation step completed within an agreed staff-time budget"],"risks":["Vendor framing driving policy","Validation that feels like policing","Inequitable burden on disabled and multilingual students"]}},"retrievedAt":"2026-10-06T03:18:37Z","enrichedAt":"2026-10-06T03:24:05Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Synchronous and oral assessments raise reasonable-adjustment demands; vague AI guidance disproportionately affects multilingual and neurodivergent students; vivas add staff workload.","procurementImplications":"Treat vendor 'learning assurance' products as unproven until independently evaluated; avoid AI-detection purchases as a primary integrity control given reported unreliability.","operatingModelImplications":"Programme-level assessment planning with consistent, explained AI expectations across modules, coordinated by teaching and learning and academic integrity offices with student input.","sourceVerification":{"openedUrl":"https://wonkhe.com/media/2026/10/Keeping-the-learning-real.pdf","referenceExcerpt":"The residual risk, as we see it, is in the area of learning validation for open assessments","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}