In June 2026, Interpol launched Operation Shams II in Tunisia, a coordinated effort to combat terrorism using artificial intelligence (AI). The operation focused on analyzing over 108,076 faces extracted from jihadist propaganda material. AI tools were used to eliminate duplicate images and ensure high quality, ultimately narrowing the dataset down to 6,362 unique, high-quality images—about 5.9% of the original collection. These images were then processed through Interpol’s Facial Recognition System (IFRS), a tool launched in 2016, with human oversight to confirm results. This method drastically reduced the workload by 94%, significantly cutting the time and resources needed for manual analysis. The operation successfully identified 126 suspected foreign terrorist fighters, showcasing the potential of combining international law enforcement collaboration with AI-assisted analysis. Interpol’s Director of Counter-Terrorism, María Carmen Muñoz González, highlighted how Operation Shams II proved the effectiveness of this hybrid approach. It is seen as a model for future international efforts aimed at identifying and tracking criminal actors, especially those involved in terrorism. This approach builds on earlier initiatives by Interpol, such as the DTECH-Light project introduced in 2019. DTECH-Light aimed to detect, extract, and analyze digital terrorist content, including identifying and locating suspects. National law enforcement agencies were asked to submit images of unnamed foreign fighters for comparison, helping to build a more comprehensive database. While Interpol has not disclosed the specific AI tools, models, or vendors used in Operation Shams II, it is likely that the system involved scripts and automation code, potentially developed using large language models. This lack of transparency does not necessarily imply the use of a fully custom-built AI system, but rather highlights the complexity of such operations and the need for discretion in sharing technical details.