Explaining and contesting automated legal decisions

Partners: Utrecht University & Tilburg University
Type: PhD
Duration: 2023—2027

As public bodies increasingly make use of automated decision-making systems (ADS), the question of how decisions can be contested becomes a pertinent one, especially considering the often ‘black box’ nature of ADS. While several legal requirements for transparency and contestability in the context of ADS have been developed, exactly how these requirements can be operationalized in practice remains to be seen. Similarly, while the field of explainable AI (xAI) has proposed many technical solutions to explain ADS, it remains unclear whether these solutions are sufficient to realize ‘contestable AI’ from a legal point of view. This PhD project therefore aims to answer the question how meaningful contestation can be operationalized with respect to the design and application of ADS in public decision making. The project will, in addition to traditional legal doctrinal research, also include a significant empirical component including interviews, focus groups and vignette studies with e.g. citizens, judges and civil servants.