Research

Sem researches the design process of public AI systems. In his PhD project, he will develop a design theory that prescribes how the process should be organised in order to prevent, mitigate, or correct citizen harms. The design theory will be based on principles of democracy, Rule of Law, and socio-technical engineering.

Public organisations are overwhelmed by the potential of AI systems for their daily practices. At the same time, they are urged to address the harms that these systems inflict on citizens. AI systems are increasingly becoming intrinsic elements of public policy. Thereby, these systems have impact on citizens' lives, sometimes leading to citizen harms such as discriminatory outcomes or health problems. Therefore, public organisations implement instruments to control their AI systems (e.g., algorithm registers or impact assessments).

Despite their efforts, public organisations struggle to tighten their grip on public AI systems. Our empirical research shows that current design practices in public organisations fall short in addressing issues emerging from AI systems. Public organisations have a narrow focus on technology, fall short in translating political debates to operational levels, conceive the design process as linear rather than iterative, and their professionals are unaware of socio-technics in AI systems. The instruments proposed by public organisations tackle some challenges by, for example, creating information flows (i.e., algorithm register). However, structural changes needed to deal with AI systems are not addressed.

Biography

Sem Nouws is a PhD researcher in the Information and Communication Technology section. Before his PhD, Sem joined the  Nationale Denktank  where he worked on a policy proposal that focuses on the ethical aspects surrounding the collection of data from individual internet users. Sem holds a master degree Complex Systems Engineering and Management from the TU Delft and a bachelor degree in Law from the Erasmus University.

Read more