Abstract
Diverse risk assessment algorithms used in policing, justice administration and criminal proceedings have contributed to promptness in decision making. Controversies have emerged regarding their enforcement, due to certain subjective weighted variables that may lead to a possible breach of human rights. This article describes the main risk assessment algorithms, emphasizing the COMPAS algorithm and determining its impact on human rights. The method used is deductive-inductive, on the basis of a documentary analysis that describes the operation of algorithms, and an inductive process used to determine if human rights are guaranteed while it is used. Results indicate that, in fact, risks exist at the time of using intelligent predictive systems, derived from a lack of transparency in the design of the algorithm and bias. The above would not allow the identification of the process used to classify the risk level and the unbiased application of the evaluation, jeopardising rights like due process, freedom and non-discrimination, among others.
Keywords
Artificial intelligence; criminal justice; algorithms; human rights