{"resourceId":"ijiet-genai-university-sports-training-2026","versions":[{"version":"external-f02fa0c0ba9d7ebe3a1de5f50f93852721b332e53a33b5fa2f2240a270570217","resource":{"id":"ijiet-genai-university-sports-training-2026","title":"IJIET mixed-methods study: GenAI university training gains with privacy and deskilling risks","organization":"Hunan Mechanical & Electrical Polytechnic / IJIET","sector":"Academic sports training / higher education AI","geography":"China (university student-athletes; transfer limits to U.S. NCAA)","publishedAt":"May 12, 2026","publicationDate":"2026-05-12","eventDate":null,"sourceName":"International Journal of Information and Education Technology","sourceLabel":"Peer-reviewed mixed-methods study (full PDF)","sourceUrl":"https://www.ijiet.org/vol16/IJIET-V16N5-2584.pdf","evidenceClass":"academic-research","outcomeClass":"mixed","topics":["developers-agents","data-security","accessibility-workforce","operating-model"],"finding":"Xiao and Guo (published 2026-05-12) randomized 120 university track/field and basketball athletes to GenAI-supported versus traditional training and interviewed 10 coaches and 15 athletes. ANCOVA showed statistically significant performance gains for the GenAI group (p<0.05). Qualitative themes praise personalization and coach efficiency but also flag black-box prescriptions, data-privacy worries about ownership and selection/commercial reuse, and deskilling/over-reliance risks; authors insist human coaches retain final authority. China polytechnic setting and 12-week horizon limit direct U.S. NCAA transfer—selected as newly relevant measured GenAI training evidence against current research/coaching AI adoption.","sledRelevance":"Empirical GenAI training outcomes plus athlete/coach privacy and autonomy concerns that U.S. athletic departments should treat as cautionary transfer evidence, not plug-and-play NCAA results.","evidence":"Full PDF: ethics approval HMEP20231003; n=120; ANCOVA p<0.05; thematic privacy/deskilling findings with verbatim athlete/coach concerns.","architectureImplications":"Integrated GenAI stack combining personalized plan generation, real-time technique feedback, and motivational support with coach override—mirrors coaching-assistant patterns entering NCAA markets.","governanceImplications":"Consent must address biometric/performance-data ownership and future selection/commercial uses; black-box plans need explainability and appeal paths.","securityPrivacyImplications":"Athletes questioned who owned performance data and how it might be used for selection or commercial purposes—directly relevant to FERPA/vendor-training clauses in U.S. deployments.","caveats":"China polytechnic sample; two sports only; 12-week intervention; not an NCAA DI compliance environment; platform is researchers' integrated system.","streamIds":["college-athletics"],"roles":{"sales":"Performance leaders want GenAI personalization proof points but should not import China polytechnic effect sizes as NCAA guarantees. Discovery should separate training-plan assistants from medical injury tools and ask how athlete data ownership is disclosed today. Bound value is a cautionary measured case that gains co-occur with privacy and deskilling risks. Use the study to set expectation that human coaches remain final authority. Unsupported claims include U.S. competitive advantage, FERPA-compliant design, or injury-prevention efficacy from this paper.","engineering":"Fit is a GenAI training-plan and feedback assistant with explicit coach override, not an autonomous return-to-play system. Architecture should log prescriptions, expose rationale, and isolate biometric stores from commercial marketing uses. Prerequisites include IRB/ethics-comparable consent language and sport-specific model validation. Constraints include non-U.S. sample and short study window. Security must answer ownership, retention, and selection-use questions athletes raised. Proof of value: blinded performance checks plus consent comprehension tests—not only p-values from another country.","delivery":"If piloting similar GenAI training tools in a U.S. department, add consent refreshers, coach explainability training, and a deskilling watch (mandatory human observation hours). Dependencies include sports medicine, compliance, and IRB/privacy office. Ownership: sports performance for training outcomes; compliance/privacy for data uses. Acceptance criteria: written ban on using GenAI outputs alone for roster selection, documented coach approval on plan changes, and athlete-readable data-use notice. Risks include over-transfer of foreign results and silent commercial reuse of athlete performance data."},"roleDetails":{"sales":{"customerProblem":"Departments want evidence that GenAI training aids work, while athletes and coaches worry about data ownership, selection use, and skill atrophy.","stakeholders":["Director of sports performance","Sport coaches","Student-athlete representatives","Privacy / compliance office"],"discoveryQuestions":["How is athlete performance-data ownership disclosed in current wearable/AI agreements?","Would GenAI outputs ever inform selection, playing time, or NIL valuation?","What human observation standards will remain mandatory?"],"valueHypothesis":"GenAI personalization can improve short-term training metrics when coaches stay final decision-makers and data uses are bounded—transfer to NCAA requires local validation.","potentialEngagement":"Bounded single-sport pilot with consent/data-ownership addendum and parallel human-coaching fidelity checks; treat foreign p-values as hypothesis only.","unsupportedClaims":["NCAA DI outcome equivalence to the China polytechnic sample","Injury-prevention or medical claims","FERPA/HIPAA compliance of the study platform"]},"engineering":{"fit":"Coach-supervised GenAI plan generation and technique feedback with explainability and consent-aware biometric storage.","architectureIntegration":"Plan generator + feedback module + audit log of prescriptions; coach approval gate; strict separation from marketing/NIL systems.","prerequisites":["Sport-specific validation dataset","Consent language covering ownership and future uses","Coach override and rationale display"],"constraints":["External study generalizability limits","12-week evidence horizon","Not a medical device"],"security":"Answer who owns performance data; block selection/commercial secondary use without fresh consent; minimize vendor training rights.","proofOfValue":["Local blinded performance comparison vs control","Athlete consent-comprehension quiz pass rate","Rate of coach overrides of AI plans"]},"delivery":{"work":"Pilot design, consent refresh, coach enablement on explainability, and monitoring for over-reliance/deskilling.","dependencies":["Privacy/IRB review","Sports medicine liaison","Athlete advisory input"],"ownership":"Sports performance owns training outcomes; privacy/compliance owns data-use enforcement; coaches own final plan authority.","skillsAdoption":"Train coaches to treat GenAI as second opinion; athletes on data rights and how to appeal plans.","governanceCheckpoints":["Pre-pilot consent and data-use sign-off","Mid-pilot deskilling/over-reliance review","End-pilot local efficacy vs foreign baseline"],"acceptanceCriteria":["No roster/selection decision based solely on GenAI output","100% AI plan changes coach-approved before athlete assignment","Athlete-readable notice covering ownership and commercial-use bans"],"risks":["Over-transfer of non-NCAA results into procurement claims","Silent secondary use of performance data","Coach skill atrophy from over-reliance"]}},"retrievedAt":"2026-09-26T03:15:35Z","enrichedAt":"2026-09-26T03:16:40Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Can augment scarce coaching attention for technique feedback while risking deskilling if observation standards erode.","procurementImplications":"Require local validation, data-ownership clauses, and bans on selection/commercial secondary use—do not treat foreign significance tests as U.S. acceptance criteria.","operatingModelImplications":"Supports augmented-coach models with explicit human-final authority rather than autonomous training autopilots.","sourceVerification":{"openedUrl":"https://www.ijiet.org/vol16/IJIET-V16N5-2584.pdf","referenceExcerpt":"Athletes and coaches raised questions about who owned the performance data and how it might be used","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}