The Politics of AI, Algorithms and Data

---

---

Abstract

Contemporary developments in the areas of machine learning (ML), artificial intelligence (AI) and the development of algorithms that can be put to work on so-called ‘big data’ represent a significant leap in digitalization processes. Indeed, it is often stated that the combination of these new computational techniques and the seemingly endless streams of data produced by human interaction with various computer interfaces is one side of a fourth industrial revolution (4IR). So far, the research focused on these developments have largely treated AI, ML and algorithms as neutral instruments that can be deployed to help and enhance human lives in general, and produce more accurate, unbiased decision making.

On the contrary, for us, this development is particularly interesting because of its anti-political ethos and its profoundly political dimensions. Collectively, we refer to these new technological instruments as Automated Systems of Governing (ASG). We argue that such systems can never be positioned beyond politics or outside of prevailing governing rationalities. Since governing is not an activity restricted to formal political bodies, we may study the effects of ASG in a wide variety of settings. Currently, our research is concentrated around the following topics:

From ethics to politics. We are interested in why and how so much of the official political debates and governing structures of ASG is focused on ethics rather than politics. We conduct theoretical as well as empirical work on the implications of this ethics focus as we develop alternative ways of conceiving ASG. In particular, we seek to examine the conditions needed to politicize these new systems, which we argue would move the discussion beyond ‘bias’.

ASG and the public sector. We are interested in how the implementation of these new systems across the public administration have organizing effects in the sense that they rearrange existing structures, produce new subjects, enable new forms of expertise and advance certain political rationalities. In particular, we are interested in the diffuse public/private arrangements that generally accompany the introduction of specific systems as they open up new spaces and venues for private interests and pose democratic questions.

The nature of ASG. We conduct theoretical work on the ontology, epistemology and ideology of ASG. When doing so, we are interested in how ASG are being assembled and what sort of performative effects they can be said to have at a general level of understanding. In this context we draw on a longer legacy of social scientific studies of governing through numbers and statistics. Additionally, we ask what kind of machines ASG can be said to be, how they produce realities and how they do certain work in contemporary capitalism.

Data protection and surveillance. We conduct policy analyses and genealogies where we trace how present discourses of competition and digitalization allows for a radical change in how surveillance is understood. We seek to show the conjunctions of political contestation that have led to a situation where far reaching mass surveillance by private firms is now accepted and even promoted as vital for growth.