Democratic Algorithms? Issues and frictions in the development of public value algorithms.
Nikolaus Poechhacker, Karl-Franzens Universität Graz
Abstract
Can an algorithm be democratic? And how can we understand algorithms not only as technical, but also as social and political phenomena? Based on ethnographic research I describe the development of a “constitutional” machine learning powered recommender system in a public broadcaster and how the implementation of the recommender system can be understood as a genuine social process. In my talk I discuss how the practical implementation of AI negotiates different social orderings with each other, enabling and limiting algorithmic power at the same time. Further, I address the question STS scholars’ role as embedded researchers in AI development projects.
Biography
Nikolaus Poechhacker is postdoctoral researcher at the Complex Social & Computational Systems research group, University of Graz. He received his PhD in Sociology from the Technical University of Munich and graduated from the University of Vienna in Science and Technology Studies (MA) and Sociology (BA), during which he also took courses in Computer Science (EC). In his work he researches the relationship between democratic institutions, law, social order, and algorithmic systems in various domains, bringing together perspectives from media theory, STS, computer science, and sociology.
Location
STS Seminar Room / Universitätsstraße 7 (staircase II / 6th floor) 1010 Vienna