Curbing Algorithmic Kafka: Embedding design processes of public algorithmic systems in a democracy and Rule of Law context

Public organisations increasingly rely on algorithmic systems for the execution of their tasks and the provision of public services. These public algorithmic systems can inflict harms on citizens by creating Kafkaesque situations. Exemplary cases of Kafkaesque algorithmic systems are the Dutch childcare allowances scandal and the Australian Robodebt scheme. Algorithmic Kafka emerges from possibilities for arbitrary conduct in the constitution of algorithmic systems. Following from the fundamental principles underlying democracy and the Rule of Law, governments should protect citizens from any arbitrary use of power by public organisations. Nonetheless, current design practices of public algorithmic systems do not provide this protection and are lacking democratic legitimacy.

This thesis provides design principles for institutional interventions that restructure design practices of public algorithmic. More specifically, these design principles prescribe how current technocratic and businesslike design practices can be shifted to practices that are embedded in a democratic and Rule of Law context. The institutional interventions facilitate public servants in coordinating the formulation of a socio-technical specification for public algorithmic systems. Furthermore, the interventions reduce arbitrary use of designerly power by strengthening the dialectic between public servants and politicians. The design practices that emerge by establishing the position of system-level designer will curb the emergence of algorithmic Kafka.