Research group AIR develops and deploys AI methods that enable new research approaches and enhance existing research practices in other disciplines.
AIR brings together researchers and research activities dedicated to applying AI in interdisciplinary research. Its work centres on three main activities:
Together, these activities create opportunities for Umeå University and its international collaborators to establish new, specialised areas of interdisciplinary research. The group has been directly involved in supervising more than 20 thesis projects.
Our core research interest lies in understanding and optimizing the process of developing AI methods for facilitating research in other disciplines. While our interests are generalist, our core activity mostly revolves around the development of AI systems capable of sensing, accounting for, or responding to psychological and social factors. We also engage in interdisciplinary research oriented towards social value (e.g., businesses, administrations, industry). Our group is constituted of likeminded researchers from numerous disciplines who are interested in these questions (e.g., education, health, industry, disaster recovery, sustainability). Our research activities span different stages of development, from initial discussions with collaborators to the integration of AI methods into established research infrastructure.
AIR core operations involve the development of software solutions for addressing practical research and society problems. Towards achieving long-term capacity building on interdisciplinary AI, we consolidate these solutions as infrastructures and research platforms. While we have several punctual contributions to a variety of applications of AI for interdisciplinary research, the following platforms stand out, which have been tested in practical research settings. Through rigorous AI-based interdisciplinary research, we have established procedures and specialised methods to address technical and practical research challenges, including procedures for establishing ethical and legal compliance when sensitive data is involved.
This platform streamlines and scales up qualitative research, including enhanced data collection (e.g., chatbots), preprocessing (transcription, cleaning and structuring), and analysis (theme identification and coding). Our infrastructure has been used in a variety of application domains, including education, health, and social services.
See as examples:
Vanhée, L., Boström, E., Lithner, J., Andersson, C., Palmberg, B., Winberg, M., Svahn, O., & Palm, T. (2026). 2CARE: LLM-Based Classifier of Imitative Versus Creative Reasoning Achieving Near-Perfect Accuracy in Think-Aloud Transcripts. IEEE Access.
Javed, R. T., Nasir, O., Borit, M., Vanhée, L., Zea, E., Gupta, S., & Qadir, J. (2022). Get out of the BAG! Silos in AI ethics education: Unsupervised topic modeling analysis of global AI curricula. Journal of Artificial Intelligence Research, 73, 933-965.
Garcia, D., Granjard, A., Vanhée, L., Berg, M., Andersson, G., Lasota, M., & Sikström, S. (2025). AI-driven analyzes of open-ended responses to assess outcomes of internet-based cognitive behavioral therapy (ICBT) in adolescents with anxiety and depression comorbidity. Journal of Affective Disorders, 381, 659-668.
This simulation platform is dedicated to streamline the development of agent-based simulations, that is representing social processes down to every individual decision (i.e., AI agents) and their interaction to support prediction, exploration, explanation, and demonstration of theories and interventions. This infrastructure has been used for facilitating interdisciplinary research in disaster recovery, urban sustainability, education, and management.
See as examples:
Edmonds, B.; Vanhée, L. (2026) A Simple Model of Anticipatory Anxiety in Students. In Social Simulation Conference
Vanhée, L.; Borit, M.; Andersson, C.; Danielski, I.; Nair, G.; Nilsson, M.; Pastrav, C.; Sjöström, J. (2026) InTrA or looking behind the scene of integration: A framework to reflect over transdisciplinary agent-based modelling with stakeholder participation, demonstrated on sustainability simulation development. In Social Simulation Conference
Lorig, F.; Vanhée, L.; and Dignum, F. (2023). Agent-Based Social Simulation for Policy Making. Human-Centered Artificial Intelligence
Beyond pure technical research, the group also actively develops analytical, organizational, and educational infrastructures for facilitating AI for education research.
See as examples:
Vanhée, L.; (2023) Organizing Transdisciplinary Research and Innovation: the Case of the Socially-Aware Artificial Intelligence Focus Area. In International Workshop on Multidisciplinary Perspectives on Human-AI Team Trust;
Vanhée, L. and Danielsson, K. and Grill, K. and Enqvist, L. and Borit, M. (2024). Hack it with EDUCHIC! Educational Hackathons and Interdisciplinary Challenges – Definitions, Principles, and Pedagogical Guidelines. European Journal of Education; 59, e12658;
Vanhée, L. and Melania B.; (2024) How Interdisciplinary Are Agent Based Models in General and Those of Human Decision in Particular? A Bibliometric Analysis. In Social Simulation Conference;
Fields: education, health, social services, psychology, foresight studies, defence, disaster recovery, energy and sustainability.
Partners: local and regional organisations, such as municipal councils and regional authorities; national partners; and international collaborators in countries including Finland, France, Norway and the Netherlands.
We offer:
● Technical expertise, resources and tools for implementing AI methods across disciplines and sectors.
● Advice on identifying the challenges and opportunities associated with using AI in interdisciplinary research, education and practical applications. We primarily work with researchers, while also supporting decision-makers and other societal stakeholders.
● BSc and MSc thesis opportunities in several disciplines, building on established expertise in computer science and cognitive science. Our supervision process provides students with tailored opportunities to contribute to active research projects and continue their work beyond the thesis, including through publications in leading venues.
Please contact us to learn more or discuss a potential collaboration!
AIR members receive funding from several research funders and programmes, including The Swedish Research Council (Vetenskapsrådet), The Wallenberg AI, Autonomous Systems and Software Program – Humanity and Society (WASP-HS), The Swedish Research Council for Sustainable Development (Formas), and The SciLifeLab & Wallenberg National Program for Data-Driven Life Science (DDLS).
