This study examines how generative AI adoption is constrained not just by individual attitudes but by structural differences across disciplines and roles. According to arXiv (AI), researchers surveyed 272 academic and professional services staff at a Russell Group university to probe whether barriers are produced by organizational ecosystems rather than only by perceived usefulness or ease of use. This matters because one-size-fits-all training will fail if institutional norms and governance are the real obstacles.
**Core claim:** Adoption barriers for generative AI are multi-level and shaped by discipline and institutional role, not just personal perceptions.
**Evidence:** The authors used multinomial logistic regression, structural equation modelling, and semantic clustering of open-ended responses from 272 staff. Non-STEM academics flagged ethical and cultural concerns—especially around academic integrity—while STEM and professional services staff emphasized governance, infrastructure, and institutional constraints.
**Institutional shift:** The paper argues universities must move beyond generalized workshops toward targeted governance, infrastructure investment, and role-specific support frameworks that address distinct epistemic norms and operational needs.
**Criticisms and limits:** The sample is from a single Russell Group university, which limits generalizability. Cross-sectional survey data and self-report also constrain causal claims. Some nuances within disciplines may be underexplored.
Insight / Analysis:
This finding is meaningful: it reframes GenAI adoption as an organizational design problem, not merely a training gap. Policymakers should prioritize policy design, technical infrastructure, and disciplinary dialogue. The study could push further by comparing multiple institutions and tracking longitudinal change, but its multi-method approach already strengthens the case for tailored strategies.
Takeaway:
If your institution treats GenAI adoption as an individual learning issue, rethink priorities. Invest in governance, role-specific support, and discipline-sensitive policies to enable responsible, effective use. Start aligning governance with pedagogy now, not after more pilots immediately.
—
**Source:** arXiv (AI)
**Original Article:** https://arxiv.org/abs/2603.27052
