Abstract:
The widespread availability of generative artificial intelligence since late 2022 has prompted higher education institutions worldwide to reconsider AI integration into teaching, research, and administration while preserving academic traditions. This study examines how university leaders at seven of Romania's leading AI-focused institutions navigate this challenge within the European regulatory context – a national perspective underrepresented in existing scholarship. Through semi-structured interviews with senior leaders, we explore support structures for AI adoption, balancing innovation with human-centered pedagogy, and future competencies needed in AI-influenced environments. Drawing on boundary-work theory, we analyse how leaders define institutional boundaries around appropriate AI use, stakeholder interests, and competing values. Findings reveal that leaders function as active mediators operating at institutional, pedagogical, and epistemic levels, negotiating AI integration through boundary-work rather than passively accepting technological change. The study contributes to scholarship on institutional governance and strategic responses to AI by foregrounding leadership perspectives, an underexplored but critical dimension of AI integration.