Enhancing intelligence on firearms trafficking in Europe: A methodological framework using open sources and LLMs

• In Event: WG Organized Crime and Criminal Networks Panel 1. Organized crime & illicit markets

Abstract

Firearms trafficking remains a significant security concert across Europe, with challenges in obtaining up-to-date intelligence due to varying data availability. Various studies have demonstrated the effectiveness of utilizing open-source data and text mining techniques in analysing criminal activities such as terrorism, human trafficking and human smuggling. Within the scope of the Horizon Europe CEASEFIRE Project, this study proposes a methodological framework to enhance intelligence on firearm trafficking by leveraging open-source data and large language models (LLMs). The framework focuses on analysing multilingual news sources across Europe to identify incidents involving firearms (seizures and selected firearm-related crimes). Through the extraction of relevant entities such as dates, locations, perpetrators, victims, and firearms used, this approach facilitates a timely and data-driven perspective of the phenomenon. Furthermore, the study integrates a risk assessment model, utilizing both incident data and official statistics, to evaluate risks and inform decision-making. This framework serves as powerful tool for law enforcement to proactively address firearms trafficking. The research highlights the potential of AI-driven methodologies in improving the intelligence landscape and in guiding responses to a growing transnational challenge.

Authors

• Marina Mancuso, Transcrime - Università Cattolica del Sacro Cuore

• Gergana Rosenova Tsakova, Università Cattolica del Sacro Cuore – Transcrime

• Deborah Manzi, Università Cattolica del Sacro Cuore – Transcrime

• Ernesto Savona, Transcrime / Università Cattolica del Sacro Cuore (Milan)