Inteligencia de fuentes abiertas (OSINT) y motivación judicial: un estándar tridimensional para la prueba penal

The use of open-source intelligence (OSINT) in criminal investigations makes it possible to locate, preserve, connect, and analyse publicly accessible digital information. When this process incorporates artificial intelligence, the apparent objectivity of its output may conceal decisions concerning data selection, authenticity, correlation, and attribution. Within this framework, the research is guided by the following question: what reasons must a court provide when admitting, assessing, or excluding an AI-assisted OSINT report to ensure that its decision is reviewable and compatible with the right of defence? It adopts a legal-dogmatic and argumentative reconstruction methodology, supplemented by cases from Argentina, Spain, and international criminal justice, and distinguishes among the source, its preservation, analytical transformation, and the resulting report. It argues that the analyst's testimony and a generic reference to technical protocols are insufficient to justify the judicial decision. The article proposes a three-dimensional standard for judicial justification comprising the legitimacy of obtaining and using the information, epistemic reliability, and effective adversarial scrutiny. It then develops a graduated response to defects: remediation, corroboration, reduced evidentiary weight, or exclusion, depending on their nature and impact.