{"resourceId":"wright-transcription-not-generation-integrity-2026","versions":[{"version":"external-ab9281bcc03778714c4073ac52ab4eb0517ef6bdb207c83e3156e785e215da83","resource":{"id":"wright-transcription-not-generation-integrity-2026","title":"Transcription ≠ generation: integrity policies must separate assistive AI format conversion from GenAI authorship","organization":"University of Exeter / International Journal for Educational Integrity","sector":"Higher education assessment integrity / disability equity","geography":"United Kingdom primary framing with international (incl. U.S.) policy references","publishedAt":"Received 19 Feb 2026; Accepted 4 August 2026; IJ EI (2026) 22:24","publicationDate":"2026-08-04","eventDate":null,"sourceName":"International Journal for Educational Integrity (Springer Open)","sourceLabel":"Wright (2026) — Transcription is not generation","sourceUrl":"https://link.springer.com/content/pdf/10.1007/s40979-026-00234-w.pdf","evidenceClass":"academic-research","outcomeClass":"cautionary","topics":["governance-procurement","accessibility-workforce","data-security"],"finding":"Conceptual/policy analysis arguing HE GenAI assessment policies often lack technical precision to distinguish generative AI (machine produces assessed intellectual content) from non-generative AI-powered tools—OCR, voice-to-text transcription, handwriting recognition—that perform clerical format conversion on student-authored content. Notes widespread faculty-facing GenAI guidelines (e.g., An et al. 2025 on top U.S. universities) and UK QAA/Russell Group guidance. Argues treating accessibility transcription as misconduct raises fairness concerns under equality law (UK Equality Act reasonable-adjustment duties; parallels to ADA/Australian DDA). Recommends explicit policy language and an evidential framework separating generation from transcription.","sledRelevance":"Pass B scrutiny complementary to 2026-09-16 Shepherd Against Prohibition: adds technical policy precision on assistive vs generative tools—critical as student-success units deploy voice/OCR supports alongside GenAI tutors. Unarchived; accepted Aug 2026.","evidence":"Source type: conceptual and policy analysis with interpretive arguments and proposed evidential framework—not an empirical outcomes study. Limitations: UK-weighted legal framing; tribunal outcomes not claimed; depends on cited secondary surveys of university guidelines; no measured false-positive rates in this paper.","architectureImplications":"Interpretation: student-success and assessment systems should tag tool-use modality (transcription/OCR vs generative) in disclosure workflows and avoid one-bit “AI used” flags that collapse assistive use into misconduct.","governanceImplications":"Source: policies need technical distinctions and equality-law-aware reasonable adjustments. Interpretation: integrity codes for AI tutoring/advising must whitelist documented assistive non-generative uses.","securityPrivacyImplications":"Transcription tools still process student audio/documents. Interpretation: apply vendor controls and disability-data sensitivity even when content is non-generative.","caveats":"Not empirical; legal characterizations are interpretive. Transfer to U.S. requires counsel review of ADA/Section 504 interplay with AI integrity rules. Accepted Aug 2026 but not previously archived.","streamIds":["student-success"],"roles":{"sales":"Sales framing: when selling GenAI integrity or proctoring/detection adjacent to student success, lead with harm reduction—policies that conflate transcription/OCR with generation can punish disabled students. Wright’s 2026 analysis urges technical precision and equality-aware adjustments. Bound offerings to policy redesign and disclosure UX that separate assistive format conversion from generative authorship; unsupported: that detectors solve this distinction.","engineering":"Engineering framing: build disclosure and evidence capture that records tool class (OCR/STT vs generative) and keeps assistive pipelines out of misconduct autoscores. Prerequisites: accessibility services integration and allowlisted assistive vendors. Constraints: UK legal framing. Proof of value: reduced false misconduct flags on documented accommodations.","delivery":"Delivery framing: joint work for integrity, disability services, and assessment offices to rewrite codes and train case handlers. Ownership: academic integrity with disability support. Risks: continued blanket bans and detector-driven discrimination."},"roleDetails":{"sales":{"customerProblem":"AI integrity rules risk treating assistive transcription/OCR as cheating, creating disability equity exposure.","stakeholders":["Academic integrity","Disability/accessibility services","General counsel","Assessment leads","Student success"],"discoveryQuestions":["Do policies define generation vs format conversion?","How are OCR/STT accommodations documented?","Are detectors used on accommodated submissions?","Who adjudicates borderline tool use?"],"valueHypothesis":"Precise tool-class policies and disclosure UX reduce wrongful misconduct findings while still governing generative authorship.","potentialEngagement":"Policy language workshop plus disclosure-form redesign and case-handler training for integrity/disability teams.","unsupportedClaims":["Empirical proof of tribunal outcomes from this paper","That banning GenAI is disability-safe","Detector accuracy guarantees fairness"]},"engineering":{"fit":"Fits assessment platforms and integrity case systems needing tool-use taxonomy fields.","architectureIntegration":"Disclosure schema with tool-class enum; allowlisted assistive connectors; human review queue before AI-misconduct flags.","prerequisites":["Policy taxonomy","Accommodation registry integration","Vendor allowlists","Audit logs"],"constraints":["Jurisdictional legal variation","Students using consumer STT","Incomplete accommodation disclosure"],"security":"Audio/documents in transcription flows need DPIA-level controls and retention limits.","proofOfValue":["Measure false-positive misconduct referrals involving assistive tools","Time-to-resolution for accommodation-related cases"]},"delivery":{"work":"Rewrite integrity codes, train investigators, and align disability services referral paths.","dependencies":["Counsel review","Faculty senate approval","Platform configuration","Student communications"],"ownership":"Academic integrity office jointly with disability services.","skillsAdoption":"Case handlers need technical literacy on OCR/STT vs GenAI; faculty need syllabus examples.","governanceCheckpoints":["Equality impact assessment","Detector-use restriction sign-off","Annual policy refresh"],"acceptanceCriteria":["Published generation-vs-transcription definitions","Disclosure form fields live","Investigator checklist updated"],"risks":["Policy lag behind tools","Inconsistent faculty enforcement","Chilling effects on accommodation use"]}},"retrievedAt":"2026-09-18T03:06:28Z","enrichedAt":"2026-09-18T03:15:52Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Source centers disability equity: non-generative AI accessibility tools should be governed as reasonable adjustments, not presumptive misconduct—directly shapes student-success accessibility staffing and advocacy.","procurementImplications":"Interpretation: RFPs for integrity/detection must require assistive-use exemptions and human review; procure campus STT/OCR that meet accessibility and privacy bars.","operatingModelImplications":"Interpretation: operating model pairs integrity adjudication with disability services before sanctioning AI-related cases involving format conversion.","sourceVerification":{"responseId":"curl-springer-wright-transcription-2026-09-17","openedUrl":"https://link.springer.com/content/pdf/10.1007/s40979-026-00234-w.pdf","referenceExcerpt":"non-generative uses of AI-powered tools, such as optical character recognition","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}