Criminal Profiling through Artificial Neural Networks.

Criminal profiling is an investigative technique which aims to reconstruct the psychological, behavioural and social profile of a potential perpetrator, based on an analysis of the crime scene and how the crime was committed. Although it is a well-established tool in forensic practice, new possibilities for its development have emerged in recent years thanks to artificial intelligence technologies, particularly artificial neural networks (ANNs). This thesis analyses how ANNs, which are inspired by the way biological neurons function, can be used in criminal profiling to identify patterns in investigative data and help create criminal profiles. Machine learning enables neural networks to process large amounts of information and detect correlations that would otherwise be overlooked, making investigative work more accurate and timely. However, integrating such tools into the investigative context raises important legal and ethical issues that cannot be overlooked. Questions arise in particular about the transparency of algorithmic decisions, the protection of personal data, the risk of bias and discrimination, and respect for fundamental individual rights such as the presumption of innocence and the right to a fair trial. The use of 'black box' tools, whose decision-making logic is not immediately understandable even to developers, requires careful consideration in terms of responsibility and human control. This paper seeks to provide a balanced overview of the operational potential of artificial neural networks in criminal profiling, as well as the regulatory and ethical considerations for their use in a manner that is compatible with the fundamental values of our legal system. The aim is to stimulate critical, multidisciplinary reflection on the use of intelligent technologies in the criminal justice sector.