Media bias is a complex and inherently interdisciplinary concept, be it in industry applications, government processes, journalism, research, or even just personal communication. Across and within such communities, different forms of bias are often defined, named, and measured in fragmented or overlapping ways, making findings difficult to compare and limiting cumulative progress. Our goal is to work toward a shared theoretical foundation that connects perspectives across disciplines while remaining useful for computational research and real-world applications.

Over the past two years, the Media Bias Group has been developing an initial conceptual framework for media bias that will serve as a starting point for discussion and validation throughout the seminar. Together, researchers from various domains as well as stakeholders from industry and journalism will examine how different forms of bias relate to one another and discuss their implications for annotation, evaluation, datasets, benchmarks, computational methods, infrastructure, governance, and policy.

The seminar is intended to be more than a one-week event. Its results will be published in a peer-reviewed paper and made openly available through the upcoming BARI (Bias Analysis Research Infrastructure). By combining a shared conceptual framework with open resources and community-driven development, we hope to establish a stronger foundation for computational media bias research and build a lasting international research community.