Background

After the International Workshop ‘Debiasing and Discrimination-Awareness in Machine Learning: Exploring Implicit Assumptions and Context’ held at KU Leuven/Belgium in 2018, ‘Deconstructing FAT’ @ ACM FAT*2020 is the second of a series of workshops that uses deconstruction to research the entanglement of computing practices and societal power relations.

Deconstruction questions the normativity of practices “by revealing the constructedness of seemingly ‘natural’ sense-making. It aims at denaturalising self-evident causalities that implicitly inform meaning constructions; it seeks absences and silenced contradictions that obscure the mechanisms sustaining discursive hegemonies and power relations” [1: 191].

Research on fairness, accountability and transparency in sociotechnical systems has built networks between researchers of different disciplines, putting an effort into translating between these disciplines and developing concepts that can work as boundary objects between them. Applying an ethnographic and deconstructive approach, we attempt to go further in this direction by using a collective method that goes beyond making different epistemic approaches accessible to one another. ‘Mind scripting’ [2, 3] — the method of deconstruction that we use in the workshop — is based in theories of discourse, ideology, memory and affect and strives to transform disciplinary knowledges by making visible their implicit normativities, sense-making, and ultimately world-making.

In this spirit, the objectives of this workshop series are twofold: to contribute to FAT research by searching for ways to reconfigure concepts and methods applied in this field, and, to generate empirical insights into the epistemic practices of overlapping areas of computing that work towards building a sociomaterial infrastructure for machine intelligence.

 

  1. Allhutter & Hofmann 2010. Deconstructive Design as an Approach to opening Trading Zones. Thinking Machines and the Philosophy of Computer Science, Hershey/New York: IGI Global, 175-192.
  2. Allhutter 2012. Mind Scripting: A Method for Deconstructive Design. Science, Technology & Human Values 37(6), 684-707.
  3. Allhutter 2010. A deconstructivist methodology for software engineering. Proceedings of the International Conference on Evaluation of Novel Approaches to Software Engineering, 207-213.