About the project

Public organisations increasingly use algorithmic and data-driven systems to support, automate, or augment decision-making. These systems can result in citizen harms, such as discrimination or unfair decision-making. Public organisations are hold responsible for these harms. Therefore, they are looking for ways to gain control over their systems. Currently, control is mostly attained by developing and implementing policy instruments. In the Netherlands, a consortium of public organisations has explored four of such instruments: an algorithm register, a governance framework, procurement conditions, and objection procedures. In this project, we assessed the instruments and provided the consortium with recommendations for improvements and identified gaps that need further research. The empirical insights from this study form the basis for our research on structuring the design process following premises of systems engineering, democracy, and the Rule of Law. In the latter part of this project, we will formulate an institutional design for the design process of public AI systems.

Outputs

• Toekomstverkeninning consortium publieke controle op algoritmes
  • The analysis as provided to the consortium
• Democratic or public control on public AI systems
  • Scientific analysis of the four policy instruments and a conceptualisation of public control on AI systems
• Dismantling digital cages: Examining design practices for public algorithmic systems

Collaborations

• Transdisciplinary Centre of Expertise for Just Public Algorithms
• Consortium publieke controle op algoritmes