Fair and efficient asylum procedures and artificial intelligence: Quo Vadis due process?

Abstract

In a context of high pressure on national asylum systems and a strive for efficiency, public authorities in Europe are increasingly exploring the potential of artificial intelligence-driven technologies in the asylum process. The use of this technology in the field of asylum is a growing but contentious topic, which raises important normative questions and concerns. In this context, this paper aims to analyse the potential implications for fair asylum procedures when artificial intelligence (AI) assists decision-making. Fair asylum procedures, or due process, are a central condition for guaranteeing the right to asylum and preventing unlawful refoulement, and overall ensuring trust in the asylum adjudication system. After revisiting the theoretical foundations of the concept of fair procedures, this paper develops a normative framework that can guide further reflection on the use of AI in asylum procedures. It thereby analyses the concepts that are key to the debate on the use of AI in decision-making: accuracy, efficiency but also participation. Then, drawing on scholarship in both political science and computer science, it explores potential challenges for the core values of fair procedures, considering both technical and non-technical challenges. This paper concludes that while AI promises efficiency gains for the administration, it identifies important challenges for accuracy and participation. On the basis of these considerations, it highlights the questions that should be asked and answered in order to protect the core values of fair asylum procedures.