Umeå University's logo

There is broad consensus that disciplinary higher education must include artificial intelligence (AI) content tailored to the needs of different domains. Such an approach to higher education will help cultivate the skills, knowledge and critical thinking necessary for the socially beneficial development and integration of AI into the routines, organisations and infrastructures that shape society. Such an approach to higher education further aligns with broader policy and legal recommendations and requirements for upskilling, reskilling, AI literacy and interdisciplinary education to enable individuals to live with, work with, contribute to developing, and navigate AI technologies.

The paper provides insights on how an empirical approach to the pressing question of educating AI broadly in higher education can be rewardingly pursued. We focus on the findings from an interdisciplinary small-scale qualitative study. The study is based on a workshop conducted in 2024 in Uppsala University within the WASP-ED programme – a nation-wide educational and interdisciplinary programme aimed at increasing the capability and capacity of Swedish universities in providing relevant and scalable education in AI.

The aim of the workshop was to lay the ground for the development of course syllabi on AI -focused curricula across disciplines and academic institutions in Swedish higher education. 13 researchers across nine academic institutions participated in the workshop. The researchers, primarily from the social sciences and humanities, are all engaged in interdisciplinary scholarship related to AI.

The paper presents the theoretical grounding of the workshop and its methodological design, and outlines the analytical steps taken to identify major pedagogical principles and topics in AI that could be developed and integrated into educational curricula. The results outline five themes that emerged during the workshop: AI, diversity and human rights; Educating on the foundations of AI; Pushing the boundaries of theories and methodology; Building better AI systems; and Educating professionals within certain disciplines – changing professions.

We compiled the themes into a blueprint of an advanced level 60-credit programme, which could be developed into a 120-credit master’s programme. The programme consists of four modules: Foundations of AI in Society; AI in the Wild; Pushing Boundaries Building better AI; and Thesis. The participants agreed on three pedagogical principles to align the programme content and courses with societal needs: Learning by doing; Real world challenges; Engaging stakeholders. We evaluated the blueprint by mapping it to the WASP-ED AI Curriculum and identified different opportunities to extend the blueprint into a master’s programme.

We suggest that the blueprint can be used by educational institutions to create courses and programmes aimed at developing disciplinary education as well as fostering multidisciplinary understanding of AI in society. Furthemore, we suggest that the methods through which the blueprint was developed offer valuable insights on how interdisciplinary education on AI can be productively developed.