The rise of generative artificial intelligence (GenAI) in higher education has transformed how students engage with learning, prompting both enthusiasm and unease. While prior research focuses on technology acceptance using quantitative models, qualitative insights into how students interpret these tools remain limited, especially in the Visegrad Group (V4). This study employed a coordinated qualitative design, utilizing 24 focus groups with students from social sciences, education, and management programs across V4 universities.

The study aimed to map student experiences onto the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) and identify necessary theoretical extensions. Inductive and axial coding was applied to analyse emerging themes. Seven thematic categories were identified: 1) AI adoption and utility; 2) output reliability concerns; 3) user skill; 4) ethical dilemmas; 5) impact on cognitive engagement; 6) the university’s role; and 7) future career implications. While several UTAUT2 constructs (e.g., performance and effort expectancy) were supported, others (e.g., social influence, facilitating conditions) were largely absent or transformed. The analysis proposes four critical theoretical extensions: Contingent Performance Trust (expanding Performance Expectancy), Epistemic Vigilance (expanding simple ease of use), Identity Negotiation (internalizing social influence), and Institutional Legitimacy (validating usage norms).

Findings suggest existing models only partially capture GenAI complexity. Students view GenAI as mandatory “career capital” yet experience deep “deskilling” anxiety. The institutional context shaped interpretations, creating an “ethical vacuum.” The study calls for universities to move from passive regulation to active scaffolding, implementing prompt literacy training and process-oriented assessment.