{"resourceId":"peck-hitl-genai-cc-advising-privacy-20260717","versions":[{"version":"external-7a715863c2362fdf022d9b0815af1a0994b730ec0913bcb4542ad9df0b61db58","resource":{"id":"peck-hitl-genai-cc-advising-privacy-20260717","title":"Ivy Tech design case: HITL GenAI advising on public data with FERPA-safe separation","organization":"Ivy Tech Community College of Indiana","sector":"Higher education / academic research","geography":"United States (community college pathway advising; transferable with limits)","publishedAt":"2026-07-17","publicationDate":"2026-07-17","eventDate":null,"sourceName":"Research Square","sourceLabel":"Preprint (doi:10.21203/rs.3.rs-10319737/v1)","sourceUrl":"https://www.researchsquare.com/article/rs-10319737/v1.pdf","evidenceClass":"academic-research","outcomeClass":"emerging","topics":["data-security","governance-procurement","operating-model","accessibility-workforce"],"finding":"Design-based case derives eight principles for community-college GenAI pathway advising: requirement groups not flat checklists; GenAI as explanation/option-framing not system of record; separate public program data from protected student records; human advisor review for consequential recommendations; plus agency, non-deficit language, visible uncertainty, and non-reductive workforce framing. Prototype can run FERPA-safe on public data alone. Why now: absent from archive and newly salient after ACC Digital Twin early-alert coverage and Pass A evidence that GenAI design/governance dominate raw access.","sledRelevance":"Gives U.S. community colleges a privacy-first blueprint for GenAI advising as campuses scale early-alert and pathway bots without waiting for outcome RCTs.","evidence":"Artifact/workflow design-based research on public curricular data; no learner-use or retention outcomes reported.","architectureImplications":"Split public catalog/requirement-group services from student SIS; GenAI explains options; advisors approve consequential paths.","governanceImplications":"Codify human-in-the-loop for high-stakes recommendations; make curricular ambiguity visible rather than deterministic.","securityPrivacyImplications":"Default to public-data advising mode; require additional permissions/governance before student-specific what-if tools.","caveats":"Single-case design study; no measured student outcomes; principles not yet trial-validated.","streamIds":["student-success"],"roles":{"sales":"For community-college advising and student-success leaders, this Ivy Tech design case shows how to field GenAI pathway help without ingesting student records—keeping GenAI as explanation layer and humans on consequential recommendations. Position discovery around FERPA-safe public-data architectures and advisor review workflows—not autonomous degree planners. Bounded engagement is an architecture/workshop pilot producing a requirement-group model and governance checklist, not a retention guarantee.","engineering":"Engineering takeaway: build requirement-group APIs on public catalog data; GenAI RAG over that corpus only; escalate to advisor tools for protected SIS joins. Prerequisites: clean public catalog, advisor UX for review, uncertainty display. Constraints: no outcome evidence yet. Security: hard separation of public vs student data planes. Proof of value: advisor acceptance testing and privacy review sign-off before SIS integration.","delivery":"Delivery work: map program requirements into flexible groups, stand up a public-data GenAI pathway explainer, define human-review gates for consequential recommendations, and train advisors on non-deficit prompts. Dependencies include registrar/catalog owners, advising leadership, and IT security. Ownership: advising AVP with information-security co-sign. Skills: curriculum modeling plus constrained GenAI prompting. Governance checkpoints: FERPA/privacy review and ambiguity-display QA. Acceptance: documented data-plane separation and advisor-reviewed consequential outputs; risks include checklist flattening and premature SIS coupling."},"roleDetails":{"sales":{"customerProblem":"Community colleges want GenAI advising at scale but face FERPA risk and fear of chatbot degree planners.","stakeholders":["advising leadership","student success","IT security/registrar","academic affairs"],"discoveryQuestions":["Can pathway Q&A run on public catalog data only?","Where must humans approve recommendations?","How are flexible requirement groups modeled today?"],"valueHypothesis":"Bounded value: deploy a public-data HITL advising explainer before any student-record GenAI joins.","potentialEngagement":"4–6 week design sprint: requirement-group model + privacy architecture review—not an outcome guarantee.","unsupportedClaims":["Retention or completion gains from this design alone","Autonomous GenAI degree planning","Transfer without local catalog modeling"]},"engineering":{"fit":"Fits as public-data advising explanation service with optional later SIS-gated what-if module.","architectureIntegration":"Catalog/requirement-group service → RAG GenAI → advisor console; no default student PII in model context.","prerequisites":["Public catalog extract","Requirement-group schema","Advisor review UI"],"constraints":["Design-only evidence","Single institution case","No learner outcome data"],"security":"Enforce public vs protected data planes; audit prompts for accidental PII.","proofOfValue":["Privacy review pass","Advisor task-time on review queue","Student comprehension of pathway options in usability tests"]},"delivery":{"work":"Model requirements, launch public-data GenAI explainer, train advisors, document escalation.","dependencies":["Registrar data","Advising staff time","Security review"],"ownership":"Advising leadership with CISO/registrar co-sign.","skillsAdoption":"Train advisors to treat GenAI as option-framing, not authority.","governanceCheckpoints":["FERPA assessment","Consequential-recommendation gate","Ambiguity display QA"],"acceptanceCriteria":["No student PII in GenAI context for v1","Human sign-off on consequential recommendations"],"risks":["Flattening flexible requirements","Premature SIS integration","Students over-trusting GenAI plans"]}},"retrievedAt":"2026-09-23T03:21:10Z","enrichedAt":"2026-09-23T03:24:00Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Non-deficit language and visible uncertainty support equitable pathway agency; workforce alignment must not collapse advising to labor demand alone.","procurementImplications":"Specify public-data-first advising GenAI with human-review requirements in RFPs; defer student-record features.","operatingModelImplications":"Operate GenAI as advising explanation capacity while advisors retain decision authority on consequential paths.","sourceVerification":{"responseId":"curl-rs-10319737-v1","openedUrl":"https://www.researchsquare.com/article/rs-10319737/v1.pdf","referenceExcerpt":"This means the advising model can function in a FERPA-safe mode using public data alone","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}