Perception of artificial intelligence-based technology in criminal profiling and crime linkage: a group study of Colombia’s judicial forces

Abstract

The Oral Accusatory Crime System (SPOA for its initials in Spanish) and Watson Machine Learning by IBM are two technological tools used by the Colombian judicial forces in criminal investigation. The first works as a database, while the second’s main function is to analyse unstructured data, such as a victim’s declaration while reporting a crime. These two tools, and similar tools that have served as predecessors have been proved to be useful for violent crimes linkage and criminal profiling of violent offenders. The aim of this paper is to identify the perception of Artificial Intelligence (AI) based technology in both processes aforementioned in members of Colombia’s judicial forces. To achieve this goal, 6 semi - structured interviews took place with crime analysts from different academic backgrounds, such as engineering, psychology, law, economics and sociology. Due to the subjects discussed throughout the interviews, and their role inside the organisation, the people interviewed will be referred to throughout pseudonyms. The results obtained from the interviews show that AI based technology such as SPOA and Watson are well - received within crime data analysts and crime investigators, but they recognize some challenges such as the fact that there are not enough qualified workers that are able to use these tools and exploit all their capabilities or that it might be underused.El Sistema Penal Oral Acusatorio (SPOA) y Watson Machine Learning, creado por IBM, son dos herramientas tecnológicas usadas por el personal judicial de la Fiscalía General de la Nación (FGN). La primera es una base de datos masiva y la segunda permite realizar un análisis de datos no estructurados, tal como lo es el relato de los hechos de un crimen violento. Estas dos herramientas, y sus predecesoras, han sido útiles para la vinculación de casos de crímenes violentos y para la perfilación criminal. El objetivo de este trabajo es identificar la percepción que tiene un grupo constituido por personas que hacen uso de las herramientas descritas anteriormente. Para alcanzar la meta establecida, se realizaron 6 entrevistas semiestructuradas a una muestra constituida por funcionarios judiciales de la FGN y un miembro de la Policía Nacional de Colombia. Dentro de los resultados se encontró que las personas entrevistadas consideran que tanto SPOA como Watson son útiles y efectivos para vincular casos de crímenes violentos y para el proceso de perfilación criminal. Sin embargo, hubo menciones a ciertas desventajas que presentan los programas, tales como su falta de organización o que, debido a su complejidad, no son usados a gran escala dentro de las organizaciones competentes.