{"resourceId":"daily-pennsylvanian-penn-rumi-writing-process-pilot-17-instructors-20261101","versions":[{"version":"external-5976951670c6227122c4c54c97b292f111d91a6c0d2ba1caa675c6af5ff16534","resource":{"id":"daily-pennsylvanian-penn-rumi-writing-process-pilot-17-instructors-20261101","title":"Penn pilots the Rumi writing-process platform with 17 instructors to make student AI use visible in take-home writing; faculty split on value and surveillance","organization":"University of Pennsylvania (Center for Excellence in Teaching, Learning, and Innovation); reported by The Daily Pennsylvanian","sector":"Higher education teaching, writing and academic integrity","geography":"United States (Philadelphia, Pennsylvania)","publishedAt":"The Daily Pennsylvanian, 1 November 2026","publicationDate":"2026-11-01","eventDate":null,"sourceName":"The Daily Pennsylvanian","sourceLabel":"Student newspaper reporting","sourceUrl":"https://www.thedp.com/article/2026/11/penn-rumi-artifical-intelligence-academic-dishonesty-prevention","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["knowledge-work","governance-procurement","data-security"],"finding":"The Daily Pennsylvanian reports (1 November 2026) that 17 Penn instructors teaching nearly 20 classes across the four undergraduate schools piloted Rumi this semester, a Google Docs-like writing workspace that lets instructors set guardrails on student AI use and see writing and revision history. Penn describes it as not an AI detection service; instructors can enable models such as ChatGPT, Gemini and Claude. CETLI's director calls it a cross-school, short-term experiment aimed at learning. A Wharton professor was unsure it would improve writing but expected it to make students think twice; a political scientist said it feels 'a little big-brothery' but gives confidence; a Middle Eastern Languages professor rejected such tools as easy to fool and associated with surveillance.","sledRelevance":"Shows institutions shifting from AI detectors to process-visibility tools for take-home writing, and the trust and surveillance trade-offs that student-success and teaching centres must manage.","evidence":"Interviews and statements from the teaching-centre director and four instructors; pilot scale stated; no outcome measures yet.","architectureImplications":"Interpretation: process-capture tools sit alongside the LMS and hold full revision histories, with optional integration of third-party models chosen by instructors.","governanceImplications":"Interpretation: pilots need stated evaluation criteria, student input and rules on how revision data may be used in misconduct cases.","securityPrivacyImplications":"Revision histories and keystroke-level process data are sensitive student records; one instructor raised surveillance concerns.","caveats":"Single-institution, one-semester pilot; no learning or integrity outcome data; vendor capabilities described only briefly.","streamIds":["student-success"],"roles":{"sales":"Sales framing: universities want to keep take-home writing without relying on unreliable AI detectors. Penn's pilot of Rumi with 17 instructors shows interest in process-visibility tools, but faculty were split, and one rejected them as easy to fool and surveillance-like. A credible conversation is about running a structured pilot with clear evaluation and student input. A bounded offer is pilot design and evaluation support. Do not claim the tool improves writing or prevents misconduct; no outcome data were reported.","engineering":"Engineering framing: process-capture platforms store detailed revision histories and can embed third-party models chosen by instructors. Fit depends on LMS integration, identity, data retention and model-routing controls. Prerequisites include data-protection review and clear rules on who can view process data. Constraints include students drafting elsewhere and easy workarounds noted by one professor. Proof of value should test whether instructors can act on process evidence fairly and whether false suspicion falls compared with detector-based referrals.","delivery":"Delivery framing: the operating owner is the teaching and learning centre, with academic integrity and IT security. Work includes recruiting instructors, defining what process data may be used for, gathering student and faculty feedback, and deciding on scale-up at term end. Acceptance should rest on measurable instructor and student experience and on agreed limits for misconduct use. Risks are perceived surveillance, workload for instructors reviewing histories, and the tool being gamed as one professor warned."},"roleDetails":{"sales":{"customerProblem":"Take-home writing is hard to trust under generative AI, and detectors are unreliable.","stakeholders":["Teaching and learning centre","Writing programme leads","Academic integrity office","Faculty pilot participants","Student government"],"discoveryQuestions":["Which courses depend most on take-home writing?","What evidence would you accept in an AI misconduct case?","How will students be consulted in a pilot?"],"valueHypothesis":"Interpretation: process visibility may preserve take-home assignments with less reliance on detectors, if students accept it.","potentialEngagement":"Design and evaluation support for a one-term writing-process tool pilot.","unsupportedClaims":["The tool improves student writing","The tool prevents AI misuse","Students broadly support process tracking"]},"engineering":{"fit":"Writing-process workspace integrated with the LMS; not a detector.","architectureIntegration":"LMS single sign-on and gradebook links; optional instructor-enabled model integrations; storage of revision histories.","prerequisites":["Data-protection impact assessment","Retention policy for revision data","LMS integration"],"constraints":["Students may draft outside the tool","Workarounds reported by faculty","Third-party model terms vary"],"security":"Revision histories are sensitive student records; restrict access and define permitted use in conduct cases.","proofOfValue":["Instructor time spent reviewing histories","Student-reported trust and burden","Referral outcomes compared with detector-based cases"]},"delivery":{"work":"Run and evaluate a cross-school writing-process pilot with clear data-use rules.","dependencies":["Faculty volunteers","Vendor contract for pilot term","Student feedback channel"],"ownership":"Teaching and learning centre.","skillsAdoption":"Instructors need guidance on reading revision histories and discussing AI use with students.","governanceCheckpoints":["Data-use rules approved before launch","Student consultation","End-of-term scale-up decision"],"acceptanceCriteria":["Proposed: student and instructor survey at term end","Proposed: documented rule on process data in conduct cases","Proposed: instructor review time measured"],"risks":["Perceived surveillance","Tool easily gamed","Extra instructor workload"]}},"retrievedAt":"2026-11-03T04:12:30Z","enrichedAt":"2026-11-03T04:14:36Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Not addressed; process-tracking tools may affect students who draft offline or use assistive technology.","procurementImplications":"Interpretation: Penn uses a short-term pilot route before school or university-wide adoption, which limits lock-in.","operatingModelImplications":"Interpretation: the teaching centre owns the pilot and feedback loop; instructors opt in course by course.","peerReview":"unknown","sourceVerification":{"openedUrl":"https://www.thedp.com/article/2026/11/penn-rumi-artifical-intelligence-academic-dishonesty-prevention","referenceExcerpt":"This semester, 17 instructors teaching nearly 20 classes across Penn’s four undergraduate schools piloted the program","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}