Teaching beyond boundaries: reflections on AI, multidisciplinarity, and institutional change in higher education

This paper examines how scholarly discourse on multidisciplinary higher education was reconfigured between the 2015–2020 baseline period and the post-2020 period, and how AI enters this evolving problem space. The study applies a multi-layered, time-comparative thematic content analysis, extended with normalised co-occurrence mapping and focused AI-entry channel analysis. The analysis is based on three data layers: an inductive pre-COVID baseline corpus (Dataset S1; 2015–2020; n = 100), a deductive post-2020 comparison corpus (Dataset S2; 2020–2025; n = 33), and a focused AI-entry corpus (Dataset S3; n = 45). Results show that COVID-19 made an existing obstacle system more visible and operationally specific rather than replacing it. Disciplinary dominance, institutional rigidity, teacher resistance, assessment difficulty and professional identity uncertainty remained central after 2020, while digital infrastructure, student motivation and new educational models became more strongly connected to the obstacle system. The relational analysis shows increased coupling among recurring obstacles, indicating a more coherent barrier system at the level of scholarly problem-framing. The AI-entry analysis shows that AI enters the discourse through student support and engagement, curriculum and boundary opening, integrity and governance, and assessment and feedback capacity. The strongest AI-entry signal is student support and engagement, while curriculum and boundary opening appears as an emerging design channel. The study contributes a transparent and replicable temporal comparison supported by open coding matrices, calculation files, figure outputs and documentation in a public dataset repository.