From the College Athletics edition of September 25, 2026
IJIET mixed-methods study: GenAI university training gains with privacy and deskilling risks
Hunan Mechanical & Electrical Polytechnic / IJIET · Academic sports training / higher education AI · China (university student-athletes; transfer limits to U.S. NCAA)
- Publisher
- International Journal of Information and Education Technology
- Original publication
- May 12, 2026
- Source retrieved
- 2026-09-26
What happened
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.
Why it matters
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 and measured results
Full PDF: ethics approval HMEP20231003; n=120; ANCOVA p<0.05; thematic privacy/deskilling findings with verbatim athlete/coach concerns.
Limitations and uncertainty
China polytechnic sample; two sports only; 12-week intervention; not an NCAA DI compliance environment; platform is researchers' integrated system.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-09-26; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- 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
- Discovery questions
- 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?
- Value hypothesis
- 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.
- Potential engagement
- Bounded single-sport pilot with consent/data-ownership addendum and parallel human-coaching fidelity checks; treat foreign p-values as hypothesis only.
- Unsupported claims
- NCAA DI outcome equivalence to the China polytechnic sample
- Injury-prevention or medical claims
- FERPA/HIPAA compliance of the study platform
Pre-sales engineering
Role takeaway
- Fit
- Coach-supervised GenAI plan generation and technique feedback with explainability and consent-aware biometric storage.
- Architecture and integration
- 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.
- Proof of value
- Local blinded performance comparison vs control
- Athlete consent-comprehension quiz pass rate
- Rate of coach overrides of AI plans
Delivery
Role takeaway
- 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.
- Skills and adoption
- Train coaches to treat GenAI as second opinion; athletes on data rights and how to appeal plans.
- Governance checkpoints
- Pre-pilot consent and data-use sign-off
- Mid-pilot deskilling/over-reliance review
- End-pilot local efficacy vs foreign baseline
- Acceptance criteria
- 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
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Integrated GenAI stack combining personalized plan generation, real-time technique feedback, and motivational support with coach override—mirrors coaching-assistant patterns entering NCAA markets.
Governance
Who approves, reviews and stays accountable for outcomes?
Consent must address biometric/performance-data ownership and future selection/commercial uses; black-box plans need explainability and appeal paths.
Security and privacy
What data, permissions and controls need testing?
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.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Can augment scarce coaching attention for technique feedback while risking deskilling if observation standards erode.
Procurement
What should contracts, pricing and exit terms secure?
Require local validation, data-ownership clauses, and bans on selection/commercial secondary use—do not treat foreign significance tests as U.S. acceptance criteria.
Operating model
Which teams own the service once it runs?
Supports augmented-coach models with explicit human-final authority rather than autonomous training autopilots.
Publication history
- 2026-09-25College Athletics · Issue 194 resources
Stable resource ID: ijiet-genai-university-sports-training-2026