Algorithmic concerns surrounding gender-based violence. The case of the VioGén system and other risk prediction systems

Abstract

In today's society, algorithms play a central role in decision-making, from information retrieval to the estimation of recidivism. In the context of gender-based violence, algorithmic risk assessment systems have become a critical tool for prevention and victim protection. This article analyzes the main strengths and criticisms of algorithmic risk assessment systems, with a special focus on the case of gender-based violence. In the first part, the main advantages and limitations of these systems are identified, according to their proponents and opponents. The second part reviews the main models and programs implemented in numerous Western countries, along with the various gender-based violence risk assessment systems operational in Spain, with a particular focus on the VioGén system. The third part discusses ethical and social proposals that have emerged concerning these systems.