Artificial intelligence has become one of the most significant transformations in contemporary criminal analysis and public security. Its capacity to process large volumes of data, identify patterns, and support decision-making has generated broad institutional expectations, but also epistemological, ethical, and institutional challenges that exceed a purely technocratic perspective. This article examines these challenges from the standpoint of complex thinking. Methodologically, the study is based on a qualitative analytical documentary review of scientific literature, institutional documents, and regulatory frameworks published between 2018 and 2025, retrieved from Scopus, Web of Science, Google Scholar, and repositories of international organizations. The analysis was structured around six thematic axes: complexity and uncertainty, the algorithmic reduction of crime, structural bias, opacity and explainability, meaningful human oversight, and institutional governance. The findings show that artificial intelligence should not be understood as a substitute for human judgment or as a neutral predictive mechanism, but as a support tool whose legitimacy depends on data quality, contextual interpretation, meaningful human oversight, and ethical governance. It is concluded that complex thinking provides a critical, relational, and responsible framework for understanding algorithmic systems in public security.
Jose Dicent (Sat,) studied this question.