The rapid expansion of artificial intelligence (AI), particularly generative AI, is reshaping teaching and learning practices in higher education across Europe and internationally. Universities increasingly adopt AI-driven systems for assessment, feedback, tutoring, and content generation, often justified through discourses of efficiency, personalization, and innovation (Azhar et al., 2025; Vieira et al., 2025). Despite this rapid uptake, recent systematic reviews show that research on AI in higher education remains dominated by application oriented and technology-acceptance approaches, with limited attention to pedagogy, disciplinary knowledge practices, and epistemological implications (Bond et al., 2024; Yan et al., 2024). This PhD-in-progress addresses this gap by examining AI integration as a pedagogical and didactical issue rather than a purely technical one. The study is situated within a European higher education context, where AI adoption intersects with long-standing didactical traditions, disciplinary autonomy, and emerging debates on responsible and human-centered AI. It adopts the premise that higher education is not reducible to information transfer or skills acquisition but involves inducting students into disciplinary ways of knowing: how knowledge is constructed, validated, interpreted, and judged within academic fields. Teaching is therefore simultaneously pedagogical and epistemological, as instructional decisions embody assumptions about what counts as legitimate knowledge and understanding. Emerging research indicates that engagement with generative AI varies substantially across disciplines, reflecting differences in epistemological commitments and pedagogical traditions (Qu et al., 2024). In science-based and technical disciplines, AI use is often framed as supporting efficiency, optimization, and procedural correctness (Kotsis, 2025). Humanities, social sciences, and professional fields report stronger concerns regarding judgement, originality, and the epistemic status of AI-generated outputs. While some case-based studies suggest that generative AI can support higher-order thinking when embedded in carefully designed pedagogical strategies (Fan et al., 2025), other research highlights risk of overreliance, weakened pedagogical relationships, and diminished human judgement when AI substitutes rather than supports teaching and learning processes (Crawford et al., 2024). Recent conceptualizations of AI as a ‘collaborative co-agent’ in education (Katsenou et al., 2025) further intensify these concerns by challenging traditional notions of epistemic authority and responsibility. Similarly, studies of AI-supported relational learning point to alternative design logics that move beyond efficiency yet remain under-theorized from a didactical and epistemological perspective (Bornemann et al., 2025). Together, these developments underline the need for a discipline-sensitive investigation of how AI enters pedagogical relationships and reconfigures authority, validation, and participation in knowledge practices. Accordingly, the study addresses three research questions: 1. How does the adoption and integration of AI tools in higher education teaching practices shape disciplinary epistemologies and knowledge-making practices? 2. How are pedagogical relationships and epistemic authority reconfigured through human-machine interaction in teaching and learning contexts? 3. What ethical-epistemic principles can guide AI implementation while preserving pedagogical integrity, human agency, and disciplinary responsibility? The theoretical framework integrates socio-constructivist learning theory (Vygotsky & Cole, 1978) with classical didactics, particularly the didactic triangle (teacher-student-content), understood as an epistemic relationship through which students are initiated into disciplinary knowledge practices (Kansanen, 2009; Klafki, 2012). To analyze how educators negotiate AI integration across disciplines, the study also draws on the Technological Pedagogical Content Knowledge (TPACK) framework as an organizing lens (Mishra & Koehler, 2006). AI adoption is thus conceptualized as a process of pedagogical and epistemological reconfiguration rather than simple technological uptake.

The study adopts a qualitative, multi-phase research design. Purposeful maximum-variation sampling will be used to recruit approximately 20 - 25 faculty members across four disciplinary clusters: STEM, humanities, social sciences, and professional programs. Within each cluster, participants will vary in their engagement with AI (enthusiastic adopters, cautious users, and critical resisters), career stage, and pedagogical orientation. This sampling strategy enables analysis of how disciplinary epistemologies shape AI integration practices. This project is currently on the first study. Interviews will explore how participants conceptualize disciplinary knowledge, current or emerging uses of AI in teaching, perceptions of student learning and judgement, and ethical concerns such as over-reliance, academic integrity, and epistemic responsibility. Document analysis will include curricula, program descriptions, learning outcomes, and assessment criteria to identify formalized epistemic expectations and how AI related practices align with or challenge them. Data will be analyzed using reflexive thematic analysis, combining deductive coding informed by the theoretical framework with inductive theme development. Cross-case analysis will be used to identify patterns and construct a typology of disciplinary AI integration practices.

The study is expected to generate three interrelated outcomes. First, it will develop a discipline-sensitive typology of AI integration that clarifies how different academic fields interpret the pedagogical and epistemological legitimacy of AI use. This typology will identify where AI aligns with disciplinary standards of knowledge and where it introduces epistemic tensions, extending existing research on disciplinary differences in generative AI engagement (Qu et al., 2024). Second, the study will offer empirically grounded insights into how human-machine interaction reshapes pedagogical relationships and epistemic authority. Rather than treating AI as inherently beneficial or detrimental, the findings will describe how different configurations of authority and judgement emerge within the didactic triangle: when teachers retain epistemic authority, when students delegate judgement to AI systems, and when AI outputs implicitly function as mediators of correctness. These findings respond to concerns about the substitution of human interaction and its implications for student development and educational relationships (Crawford et al., 2024). Third, the project will articulate ethical-epistemic principles for AI implementation grounded in disciplinary contexts and everyday teaching practice. These principles will provide guidance on preserving student agencies, aligning assessment with disciplinary epistemic goals, and supporting pedagogically responsible AI use. The findings are intended to contribute to European debates on AI in higher education by offering pedagogically grounded insights applicable across diverse national systems and institutional contexts (Bond et al., 2024; Yan et al., 2024).

Azhar, A., Parveen, K., & Saleem, M. (2025). Artificial intelligence for enhancing teaching and learning practices in higher education. https://doi.org/10.63468/sshrr.102 Bond, M., Khosravi, H., De Laat, M., Bergdahl, N., Negrea, V., Oxley, E., Pham, P., Chong, S. W., & Siemens, G. (2024). A meta-systematic review of artificial intelligence in higher education: A call for increased ethics, collaboration, and rigour. International Journal of Educational Technology in Higher Education, 21(1), 4. https://doi.org/10.1186/s41239-023- 00436-z Bornemann, M., Hartlieb, E., & Rausch, A. (2025). AI-supported learning for relational capital development: A case study in higher education. Proceedings of the European Conference on Knowledge Management, 26(1). https://doi.org/10.34190/eckm.26.1.3998 Crawford, J., Allen, K.-A., Pani, B., & Cowling, M. (2024). When artificial intelligence substitutes humans in higher education: The cost of loneliness, student success, and retention. Studies in Higher Education, 49(5), 883–897. https://doi.org/10.1080/03075079.2024.2326956 Fan, Y., Wang, Z., Chawla, N. V., & Zheng, L. (2025). Developing students’ higher-order thinking skills with generative AI: Insights and strategies from a case study. In Advances in Computational Intelligence and Robotics. https://doi.org/10.4018/979-8-3693-7332-3.ch024 Gunder, A., Kohnke, L., & Moorhouse, B. L. (2025). Bridging borders: Global faculty learning communities on AI for transforming teaching and learning. Ubiquity Proceedings. https://doi.org/10.5334/uproc.185 Kansanen, P. (2009). The curious affair of pedagogical content knowledge. Orbis Scholae, 3(2), 5–18. Katsenou, A., Filippidis, S. K., & Tzafilkou, K. (2025). Beyond assistance: Embracing AI as a collaborative co-agent in education. Education Sciences, 15(8), 1006. https://doi.org/10.3390/educsci15081006 Klafki, W. (2012). Didaktik analysis as the core of preparation of instruction. In Teaching as a Reflective Practice (pp. 139–159). Routledge. Kotsis, K. (2025). Integrating artificial intelligence into higher education curricula: Challenges and opportunities for science-based programs. https://doi.org/10.54660/ijaiet.2025.6.1.04-12 Mishra, P., & Koehler, M. J. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. Teachers College Record, 108(6), 1017–1054. Qu, Y., Tan, M. X. Y., & Wang, J. (2024). Disciplinary differences in undergraduate students’ engagement with generative artificial intelligence. Smart Learning Environments, 11(1), 51. https://doi.org/10.1186/s40561-024-00341-6 Vieira, C., Figueiredo, M., & Pereira, A. (2025). Generative artificial intelligence in higher education: Challenges, opportunities and pedagogical implications. RTIC – Revista de Tecnologias, Informação e Comunicação. https://doi.org/10.55267/rtic/16675 Vygotsky, L. S., & Cole, M. (1978). Mind in society: Development of higher psychological processes. Harvard University Press. Yan, L., Sha, L., Zhao, L., Martinez-Maldonado, R., Chen, G., Li, X., Jin, Y., & Gašević, D. (2024). Practical and ethical challenges of large language models in education. British Journal of Educational Technology, 55(1), 90–112. https://doi.org/10.1111/bjet.13366