Algorithmic decision making (ADM) systems are increasingly used in different parts of public life to decide together with humans (or all alone). In the US, the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) is for instance used to assess a defendant’s risk of committing more crimes. In Poland, an algorithmic scoring system is used to assess the “employment potential” of job seekers and influences the possibilities for this person to benefit from active labour market policies. Similarly, predictive policing is based on algorithms that scan big data to carve out correlational patterns and leading to decisions where police patrols are strengthened.
Whereas ADM-systems are all over the place, political science has not started to seriously engage with these developments. To date, most of the early studies on the politics of algorithms cover the “input”-side of the political system and ask, for instance, whether “filter bubbles” exist and whether ADM-systems used in social networks affect public opinion and, in consequence, influence the outcomes of elections (see for instance here). Almost nothing has been published on the “output”-side of the political system, namely whether and how policies are affected by ADM-systems.
This year, two new research projects at the professorship will start to change this. A first two-year project (“FairAndGoodADM”, financed by the German Ministry of Higher Education and Research (BMBF), will explore how three different ADM-systems that are already at work (social policy, higher education policy, penal policy), have been implemented and how they are regulated. We will work in close collaboration with the “Algorithm Accountability Lab” at the TUK (Prof. Zweig) and the Chair of Philosophy (Prof. Joisten).
The second four-year project “Deciding about, by and together with algorithmic decision-making systems” is financed by the Volkswagen Foundation and focusses on the criminal justice sector more specifically. It is an international collaborative endeavour and involves computer scientist Prof. Zweig (TUK), the legal scholar Prof. Schulz (Hamburg), the legal philosopher Prof. Yeung (Birmingham) and the economic psychologist Prof. Achtziger (Friedrichshafen). In this project, we will aim at establishing a dataset on ADM-systems that are used in the criminal justice system in all 50 US states. The dataset, which provides details on the respective ADM-systems, will then be analyzed quantitatively to carve out dynamics of policy diffusion, political influence and so on.
Two great new paths of research in the growing field of the politics of digitalization that we are happy to begin this year.
Georg Wenzelburger
[photo by Thomas Koziel, provided by KLUFOS, University of Kaiserslautern]