Conference Program

The article analyses contemporary narratives circulating in Danish universities guidelines, reports, and teacher training courses holding the promises of Generative AI (genAI) for knowledge systems automation, problematising the implications of these promises for the future of Higher Education. It draws on the concept of infrastructure theorised by Science & Technology Studies scholars, which refers tolayered, imbricated sociotechnical assemblages built into a system to provide for its smooth functioning – that is, an invisible ‘system of substrates’ (Star, 1999). The article seeks to expand this definition of infrastructure towards the inclusion of a more-than-human perspective in its conceptualisation. Being attentive to such a ‘non-anthropocentric ontology of infrastructures’ (Barua, 2021), in fact, can provide a relational analytic for grasping how infrastructures furnish ‘substrates’ for the university digitalisation agenda, and more specifically on what grounds the narrative about knowledge automation has been constructed.

Our analysis of more-than-human infrastructures includes: (i) public–private actors’ joint collaborations, where the lobbying of Big Tech, edtech companies, and data-extractive businesses is foregrounded; (ii) space-times of new academic practices with genAI (e.g. learning management systems with automated platform functionalities, computer labs where AI is naturalised as a ‘tool’ to speed up students’ academic work, or courses for teachers’ professional development where genAI is anthropomorphised as ‘sparring partner’); (iii) affective atmospheres generated by institutional expectations of upskilling for teachers and students alike; (iv) informal relational systems supporting the functioning of formal governance mechanisms (e.g. cultural norms and values, taken-for-granted assumptions, meanings, scripts, status differentiators, incentives, rewards, etc.); and (v) ‘field configuring events’ (Hinings et al., 2017), such as seminars, conferences, festivals, and courses on ‘responsible’ uses of genAI.

The analysis shows how the recent transition of Danish universities toward general acceptance of genAI tools for any purpose, including academic exams, can be understood as the last frontier of more-than-human infrastructural work that has established, also retrospectively, institutionalised techno-visions supporting the need for digitalisation – and increasingly automation – of knowledge practices. These results are put in dialogue with broader analyses of how ‘computing the incomputable’ (Parisi, 2016) has historically transformed the grounds of how both knowledge and cognition are conceived, making invisible the probabilistic nature of machine learning ‘reasoning’ and the selection of certain epistemologies over others. In conclusion, we argue that this more-than-human infrastructuring of the digitalisation agenda has contributed to conceiving academic knowledge as a set of finite outputs, learning as an efficient and linear process driven by the logic of speed (Atkinson & Flanagan, 2024), and research practices as optimised patterns’ extrapolation, anticipation, and modelling (Grimaldi, 2025). These aspects urge us to reflect on the need to reconfigure university practices for sustaining democratic futures.