{"resourceId":"educause-ai-work-survey-2026","versions":[{"version":"external-ec9a49a71eb71c47a364989330f7fe2550fef7f91897673968396ecc36b02b90","resource":{"id":"educause-ai-work-survey-2026","title":"Workforce survey exposes gaps between AI use, policy awareness and ROI measurement","organization":"EDUCAUSE with AIR, NACUBO and CUPA-HR","sector":"Higher education workforce","geography":"Predominantly United States; includes international respondents","publishedAt":"January 12, 2026","publicationDate":"2026-01-12","eventDate":null,"sourceName":"EDUCAUSE: The Impact of AI on Work in Higher Education","sourceLabel":"Association research; respondent perceptions, with disclosed Zoom sponsorship and no sponsor editorial influence","sourceUrl":"https://www.educause.edu/research/2026/the-impact-of-ai-on-work-in-higher-education","evidenceClass":"public-sector-association","outcomeClass":"mixed","topics":["knowledge-work","data-security","governance-procurement","accessibility-workforce","operating-model"],"finding":"Respondents report widespread work-related AI use alongside limited awareness of guidance and limited institutional ROI measurement.","sledRelevance":"Interpretation: use as a diagnostic starting point for campus administrative and IT work; validate conditions locally before extrapolation.","evidence":"Email survey, September 29–October 13, 2025: 1,960 qualifying responses. Reported recent AI use was 94%; policy awareness 54%; institutional ROI measurement 13%. These are respondent reports, not an institutional census or causal productivity estimates. Methods and demographic tables were inspected.","architectureImplications":"Interpretation: pair approved-tool access with a workflow inventory and data-flow map. Survey responses do not select cloud, on-premises or hybrid hosting.","governanceImplications":"Interpretation: test whether employees can apply guidance to actual tasks, separately from whether a policy document exists.","securityPrivacyImplications":"Interpretation: examine unapproved channels through a confidential staff inventory, then review sensitive-data handling and retention. Avoid treating aggregate survey responses as evidence of a local breach.","caveats":"Constrained sampling and response bias; question denominators vary. No experimental baseline, verified financial return or campus-specific diagnosis. Sponsorship is disclosed. The survey covers AI beyond generative AI.","streamIds":["campus-operations"],"roles":{"sales":"Interpretation: for CIO, HR and finance discussions, use the survey's measurement gap to ask how the campus currently knows whether administrative AI is useful. Which tasks have reliable baselines, who reviews errors, and can staff explain permitted data use? A bounded engagement could establish a workflow inventory and an evaluation plan for one service. The value hypothesis is better investment decisions and fewer unexamined exposures, subject to local evidence. Do not turn respondent percentages into campus penetration estimates, a sales opportunity list, guaranteed savings or evidence that any particular product is required.","engineering":"Interpretation: begin with a low-risk document or service-desk workflow and representative sanitized inputs. Map source permissions, retrieval boundaries and output destinations before selecting an architecture. Prerequisites include an accountable data owner and an approved test set. Run equivalent tasks with the current process and the proposed assistance, including review time and intentionally misleading inputs. Measure factual errors and unauthorized disclosure attempts as well as elapsed time. The proposed proof should reveal whether quality survives acceleration; survey enthusiasm cannot substitute for this test. Keep integrations read-only until the owner accepts evidence for expanded permissions.","delivery":"Interpretation: assign an administrative service owner, an IT implementation lead and a finance reviewer. Establish the baseline, deliver accessible task training, and record support and correction work during the trial. Dependencies include staff availability and agreement on what successful service means. Hold checkpoints before data access and before renewal. Proposed acceptance requires a complete cost ledger, documented quality assessment and staff demonstrations of correct escalation on sensitive cases. These are proposed measures, not reported results. Risks include self-selection of enthusiasts, hidden review work and pressure to convert perceived time savings into premature staffing reductions."},"retrievedAt":"2026-09-08T03:05:35Z","enrichedAt":"2026-09-08T03:07:06Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Interpretation: budget accessible training and review effort as paid work; involve staff representatives before changing job expectations.","procurementImplications":"Interpretation: require a measured renewal decision incorporating license, support, integration and review costs.","operatingModelImplications":"Interpretation: give service owners responsibility for quality and finance responsibility for benefit validation; shared governance should resolve conflicts.","updateExplanation":"New to the full archive. Older context is newly relevant to the September campus access initiatives assessed in this edition and fills the previous edition's workforce-measurement gap; not a claim of publication since the last run.","sourceVerification":{"openedUrl":"https://www.educause.edu/research/2026/the-impact-of-ai-on-work-in-higher-education","referenceExcerpt":"This research is limited by standard survey limitations such as constrained sample frame and response biases.","promptVersion":"sled-research-v3.1","model":null,"basis":"agent-reported inspection"}}}]}