Trust and provenance
How can people recognise trustworthy AI-supported content without either rejecting AI blindly or relying on it uncritically?
The Norwegian Computational Behaviour & AI Lab is an interdisciplinary research initiative led by Professor Christoph Trattner at the University of Bergen. It connects researchers, students and external partners studying how recommender systems, generative AI and digital platforms influence what people see, choose, believe and trust.
Researchers combine computational modelling with user studies, experiments and field collaborations. The focus is how AI affects decisions and behaviour in media, health and digital platforms.
Current directions
Questions for collaboration
News people tend to avoid
How can recommender systems broaden attention and curiosity instead of simply reinforcing existing interests?
Editorial judgement
How should AI support editors and journalists while preserving professional judgement, accountability and control?
Food and health decisions
How can personalised systems encourage healthier and more sustainable choices while avoiding manipulation and over-reliance?
Research domains
Fields we work across
The work connects media and democracy, health and food, and recommender systems.
Media, news & democracy
Responsible news recommendation, AI-supported journalism, misinformation, provenance, usable transparency, selective exposure and news avoidance.
Health, food & behaviour
Food and health recommender systems, digital nudging, cross-cultural dietary behaviour, sustainability and the quality of AI-generated advice.
Recommenders & digital platforms
Personalisation, popularity bias, long-term user experience, platform behaviour and human–AI interaction beyond short-term engagement metrics.
Methods
How we study behaviour and AI
The initiative combines technical development with behavioural and societal evaluation.
Computational modelling
Recommender systems, machine learning, simulations and data-driven models of user behaviour.
User studies & experiments
Controlled studies, online experiments and longitudinal evaluations of behaviour, trust and choice.
Human-centred prototypes
Interfaces, explanations and decision-support tools designed around users and professional practices.
Partner-based field research
Living-lab studies and evaluations with media, technology, health and public-sector organisations.
Research environment
Connections to groups and projects
The Lab provides a thematic home across complementary research and innovation environments.
DARS research group →
Foundational and applied research on recommender systems, personalisation and user modelling.
SFI MediaFutures →
Large-scale research–industry collaboration and real-world experimentation in responsible media technology.
Interdisciplinary projects →
Collaborations across information science, media studies, psychology, health research, HCI and data science.
Selected scholarship
Selected papers across the themes
Eight papers representing these shared research interests, from news and recommendation to health and food.
Responsible AI, media & democracy
View figure ↗C2PA Provenance Labels Increase Trust in Digital News Platforms Across Western Countries
Cross-country evidence on how content provenance affects trust in news platforms.
View figure ↗Increasing Editor Trust in News Personalization Systems with Fact-checked Large Language Models
Examines how evidence-grounded explanations can support editorial oversight of personalisation.
View figure ↗Leveraging Professional Ethics for Responsible AI: Applying AI Techniques to Journalism
Shows how professional journalism ethics can become practical design constraints for responsible AI.
View figure ↗Responsible Media Technology and AI: Challenges and Research Directions
A research agenda for responsible AI across news production, recommendation and media use.
Recommender systems & human behaviour
View figure ↗Examining the Merits of Feature-Specific Similarity Functions in the News Domain Using Human Judgments
Connects algorithmic similarity measures with how people actually judge related news content.
View figure ↗A Survey on Popularity Bias in Recommender Systems
Systematises how popularity bias emerges, how it affects stakeholders and how it can be evaluated and mitigated.
Health, food & behavioural AI
View figure ↗Supporting Healthier Food Choices Through AI-Tailored Advice: A Research Agenda
Outlines a human-centred agenda for personalised nutritional advice and responsible behaviour change.
View figure ↗Healthiness and Environmental Impact of Dinner Recipes Vary Widely Across Developed Countries
A comparison of the nutritional quality and environmental impact of 600 dinner recipes from Norway, the UK and the United States.
Collaboration and supervision
The initiative welcomes research collaboration with academic, industry and public-sector partners, as well as strong student projects aligned with its research themes.