{"resourceId":"ihe-opinion-science-of-learning-guide-ai-use-iterative-studies-transfer-20261022","versions":[{"version":"external-728eb453943c3777b8868fad6314cc1e1cba12fb8cb1c6fb37a4ea2677921df0","resource":{"id":"ihe-opinion-science-of-learning-guide-ai-use-iterative-studies-transfer-20261022","title":"Inside Higher Ed opinion: APA and Springfield College learning scientists argue the science of learning should guide AI use, and call for small iterative studies testing transfer without AI","organization":"American Psychological Association; Springfield College (authors)","sector":"Higher education teaching and learning design for generative AI","geography":"United States","publishedAt":"Inside Higher Ed (Views), 22 October 2026","publicationDate":"2026-10-22","eventDate":null,"sourceName":"Inside Higher Ed","sourceLabel":"Opinion essay by learning-science leaders","sourceUrl":"https://www.insidehighered.com/opinion/views/2026/10/22/science-learning-should-guide-ai-use-decisions-opinion","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["knowledge-work","governance-procurement","operating-model"],"finding":"In a 22 October Inside Higher Ed opinion essay, Beth M. Schwartz (APA senior director of teaching and learning) and Christopher Hakala (Springfield College) accept that evidence on whether generative AI improves student learning remains limited and that some studies suggest unguided use could undermine it. They argue that the established science of learning still gives a framework: learning requires targeted cognitive effort, so the question is whether a specific AI use strengthens or replaces that effort. They propose that faculty start from a learning goal and ask which cognitive work must remain with the student and whether students can demonstrate learning independently. They recommend that institutions support smaller, iterative studies of specific AI uses, comparing student work before and after AI assistance, testing transfer without AI and comparing results across disciplines and student populations.","sledRelevance":"Gives student-success leaders an evaluation frame for AI pilots that centres independent transfer and subgroup differences rather than tool adoption, relevant given recent mixed course-level trial results.","evidence":"Expert opinion grounded in prior learning-science literature; no new data.","architectureImplications":"Limited technical relevance. Interpretation: evaluation designs that compare work before and after AI assistance require tools that can capture drafts and AI interactions with consent.","governanceImplications":"Recommends institutions fund small iterative studies and judge AI uses against learning goals, shifting governance from tool approval to use-case evaluation.","securityPrivacyImplications":"Not addressed. Interpretation: before-and-after work comparisons involve student data that needs research and privacy review.","caveats":"Opinion, not research; the authors acknowledge the AI-specific evidence base is limited. Recommendations are not tested here.","streamIds":["student-success"],"roles":{"sales":"Sales framing: learning scientists are telling institutions to judge AI by whether a specific use builds or replaces student thinking. Use it to discover whether a campus evaluates AI pilots by transfer without AI and by student subgroup, or only by adoption. A bounded engagement could design a small iterative evaluation for one AI use. Do not present this as evidence that any tool works: it is an opinion essay without new data.","engineering":"Engineering framing: direct technical relevance is limited, but the proposed evaluations need instrumentation. Determine whether the learning platform can capture drafts before and after AI assistance, log AI interactions with consent, and support AI-free assessment conditions for transfer tests. Validate data handling under research and privacy review, and check that subgroup comparisons are feasible with available student data and sample sizes per course.","delivery":"Delivery framing: turn the essay into an evaluation routine. For each AI use, record the learning goal, the cognitive work that must stay with the student and an AI-free transfer check. Run small, term-length studies with teaching-centre and institutional-research support, and compare results by discipline and student group. Report findings to faculty and decide whether to scale, revise or stop each use."},"roleDetails":{"sales":{"customerProblem":"Institutions are investing in AI for learning without a practical way to judge which uses help students learn.","stakeholders":["Teaching and learning centre","Institutional research","Faculty senate or curriculum committee","Provost","IRB or research office"],"discoveryQuestions":["How are AI pilots evaluated today?","Do you test whether students can perform without AI?","Do you compare results by student group?","Who funds small evaluation studies?"],"valueHypothesis":"Interpretation: use-case-level evaluation could help institutions scale AI uses that support learning and stop those that replace it.","potentialEngagement":"A bounded evaluation design for one or two AI uses: learning goals, AI-free transfer measures, subgroup analysis and a one-term reporting cycle.","unsupportedClaims":["That any specific AI tool improves learning","That learning-science principles alone guarantee results","Effect sizes for AI uses"]},"engineering":{"fit":"Fits institutions running AI pilots in courses that can add evaluation instrumentation.","architectureIntegration":"LMS and AI tool logging to capture drafts and interactions, plus AI-free assessment settings.","prerequisites":["Consent and research approval","Draft and interaction capture","AI-free assessment capability","Access to subgroup data"],"constraints":["Opinion piece without tested methods","Small studies may lack statistical power","Privacy limits on interaction data"],"security":"Handle captured drafts and AI logs as student records under research and privacy review.","proofOfValue":["Pilot one before-and-after comparison","Run an AI-free transfer task","Check subgroup sample sizes","Faculty review of design questions for one course"]},"delivery":{"work":"Set up an iterative evaluation routine for course-level AI uses.","dependencies":["Faculty partners","Institutional research support","Research approval","LMS data access"],"ownership":"Teaching and learning centre with institutional research.","skillsAdoption":"Faculty development on mapping learning goals to permitted AI uses and designing transfer checks.","governanceCheckpoints":["Research and privacy approval","Learning-goal mapping before pilot","End-of-term results review","Scale, revise or stop decision"],"acceptanceCriteria":["Proposed: each piloted AI use documents a learning goal and AI-free transfer measure","Proposed: results reported by discipline and student group where sample allows","Proposed: a scale, revise or stop decision recorded each term"],"risks":["Underpowered studies","Faculty workload","Selection bias in pilot courses","Results ignored in procurement decisions"]}},"retrievedAt":"2026-10-23T03:15:44Z","enrichedAt":"2026-10-23T03:19:48Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Calls for comparing results across student populations and expertise levels, which supports equity analysis of AI uses.","procurementImplications":"Interpretation: purchase decisions could require vendors to support use-case-level evaluation and transfer testing.","operatingModelImplications":"Places design decisions with faculty and evaluation with teaching centres and institutional research.","sourceVerification":{"openedUrl":"https://www.insidehighered.com/opinion/views/2026/10/22/science-learning-should-guide-ai-use-decisions-opinion","referenceExcerpt":"Institutions should support smaller, iterative studies of specific AI uses.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}