Knowledge in a Union
Staffan Edling, Lund University

How do political organisations construct versions of the society they act in? How do social science theory, quantitative methods and the epistemic objects of non-academic political research interact with the political interests of the organisation? This presentation is about plans for an ethnographic project on the Swedish trade union confederation LO and their offices of inquiry and investigation. LO is an important centre for the production of knowledge in the Swedish labour movement, historically as well as today. Social science analyses of epistemic policy actors have had a tendency to reduce the organisations to political actors that work through conventionally social or political means (e.g. networking, lobbying), and ignore the epistemic content of reports written by organisations. Through applying an STS perspective on political research and investigations, I hope to produce an account of how versions of social/political things are brought into being and disseminated in a political context. In order to account for how heterogeneous things like the labour market, LO’s congress and leadership and statistical methods interact, I intend to use a material semiotic framework that makes few a priori assumptions about the characteristics of actors. The main theoretical inspirations for the project at this stage are works normally placed within actor-network theory (mainly Bruno Latour and John Law) and STS writers with an affinity to ANT (e.g. Annemarie Mol and Steve Woolgar). I will also draw on sociological literature that is compatible with such a framework, such as ethnomethodology and institutional ethnography. The planned means of conducting the study is an ethnography centred around the head office of the Swedish trade union confederation LO, involving active participation in the research conducted by the organisation.


Machine learners: Towards a data-driven Norwegian Labour and Welfare Administration
Lisa Reutter, Norwegian University of Science and Technology

This paper offers an ethnographic encounter with the Norwegian Labour and Welfare Administration’s (NAV) early attempts to explore its data-driven future. The Norwegian public sector is in a pioneer mood. Platforms, big data and machine learning signal promises of a profound digital transformation, that will impact all aspects of the welfare state and the services it provides to citizens and society. NAV is at the heart of the welfare state, managing one third of the state budget and delivering more than 60 services. The newly established NAV AI lab is now set on a quest to explore and employ data utilization through machine learning. I build the findings presented in the paper on an ethnographic study of this lab, inspired by early laboratory studies. By deconstructing the formalized step-by-step work model of the AI lab, this paper explores how its work is imagined to change the welfare state. Moreover, this paper examines how the data scientists are reconstructing and configurating data practices in the attempt to achieve a data-driven future. Although some politicians hope to find a solution to sustain the welfare state in the alleged data gold mine, the data scientists set on the quest to do so, seem to struggle to align new data practices with democratic values and existing infrastructures. In these first observations, the concept of the data-driven welfare state itself appears as a black box to the AI lab. The data scientists are also constantly forced to consider the future social power of the systems they create, facing a dystopian imagery surrounding what I theorize as a socio-technical assemblage. These early encounters draw attention to an ambiguity and uncertainty tied to the overall configuration of data practices. As such, they open a space for STS scholars to reflect on data practice reconstruction and purposeful interventions.


Data rush – health data policies and strategies in Denmark and Finland as examples of data market imaginaries
Heta Tarkkala, Aaro Tupasela and Karoliina Snell, University of Helsinki

The Nordic countries have maintained a unique place within the European and global health data market. They have extensive nationally maintained and centralized health data records, as well as numerous biobanks where data from individuals can be connected based on personal identification numbers. Frank (2000), for example, has suggested that in such countries the entire population becomes a study cohort. Much of this phenomenon can be attributed to the emergence and development of the Nordic welfare state system, where Nordic countries sought to collect large amounts of population data in order to guide decision-making and help improve the health and living conditions of the population. More recently, however, the so-called Nordic gold mine of data is being re-imagined in a wholly other context, where data and its ever-increasing logic of accumulation is seen as a driver for private business development and intervention. Our presentation examines and compares the development of policies and strategies for health data markets in Denmark and Finland. We pay attention to the way emerging health data markets provide insight into how a broad range of different data sources ranging from hospital records and pharmacy prescriptions to biobank sample data are brought together under a supposed harmonious national data infrastructure, which enables ‘full-scale utilisation’ of health and welfare data. Moreover, we address the challenges related to these data markets and the emerging data divides between the ones who collect, store and mine data and those who are the targets of this collection.