Jan 23 – Jan 28, 2028
The Dagstuhl Seminar Towards Standardizing Media Bias for Computational Research is an international event dedicated to advancing the conceptual foundations of computational media bias analysis. It brings together leading stakeholders from computer science, journalism, political science, psychology, AI policy, and industry to discuss how media bias should be defined, categorized, annotated, and evaluated across different disciplines and media formats.
By interdisciplinary exchange, the seminar aims to improve conceptual clarity, support more comparable and reproducible methologies, and especially strengthen international collaboration around computational approaches to media bias analysis, evaluation, and benchmarking.
Over two years of preparation, the team around the Media Bias Group has developed an initial ontology of media bias, which will serve as a starting point for discussion alongside other core topics. Through short overview talks, plenary discussions, and collaborative working groups, participants will refine a shared conceptual framework for media bias and explore its implications for annotation, evaluation, datasets, benchmark design, and computational methods.
Beyond creating a major forum for collaboration among bias-related stakeholders, we plan to publish the seminar’s results as a research paper. We will also make the outcomes freely available and directly applicable at scale through the BARI Bias Analysis Research Infrastructure. BARI will provide structured access to the consolidated framework and support the seamless classification of different bias types. As an early kickoff for this infrastructure, we have secured approximately €1 million in funding.
Coordinator, Media Bias Group, National Institute of Informatics, Tokyo
Director and Professor, Information and Society Research Division, National Institute of Informatics; Director, Global Research Center for Synthetic Media; Professor, University of Tokyo
Project Lead, Data Innovation Lab; Senior Advisor on International Digital Policy
Professor, Political Science, University of Zurich
Professor of Data Science and ARC Future Fellow, University of Queensland; Dieter Schwarz Fellow, Technical University of Munich
Media bias influences public opinion, political discourse, and the information ecosystems that shape modern societies. Despite growing research activity, there is still no broadly accepted framework describing what media bias is, how it should be measured, or how computational systems should evaluate it.
This Dagstuhl Seminar aims to bring together experts from diverse disciplines to discuss these challenges from multiple perspectives. Throughout the week, participants will collaboratively examine existing concepts, identify open questions, and explore opportunities for future cooperation.
Beyond the seminar itself, we hope to strengthen the international research community working on media bias and lay foundations for future collaborations, shared resources, and interdisciplinary initiatives.