Predictive algorithms and the prevention of femicide: possibilities and limits within the criminal justice system

Abstract

 This article analyzes the use of predictive algorithms in the criminal justice system as a tool for the prevention of femicide, in light of Criminal Law, Criminal Procedure, and Fundamental Rights. It starts from the empirical observation of increasing femicide rates in Brazil, highlighting the need for more sophisticated preventive strategies. In this context, the study examines the potential of artificial intelligence and machine learning in identifying risk patterns in situations of domestic violence, as well as their ethical, legal, and institutional limitations. A qualitative methodology is adopted, including a recent bibliographic review (2021–2026), normative analysis, and an interdisciplinary dialogue with data science. The study concludes that, although predictive algorithms represent a promising tool for preventive public policies, their use requires strict parameters of transparency, democratic oversight, and respect for fundamental rights.