Interpol says AI helped identify 126 suspected fighters from over 100,000 images

Interpol says an operation it coordinated in Tunisia in June 2026 used artificial intelligence to process more than 100,000 images from jihadist propaganda and produce matches for 126 suspected foreign terrorist fighters.

Programmed AI agents and AI-generated scripts removed duplicates and screened out poor-quality faces. The process reduced 108,076 images to 6,362 unique, identifiable images, which were then checked with the Interpol Facial Recognition System (IFRS), launched in 2016.

Interpol said the remaining set represented about 5.9% of the original collection and eliminated 94% of the images that investigators would otherwise have had to process manually. The 126 people were identified on paper as potential targets and would still need to be located and investigated.

Interpol Director of Counter-Terrorism María Carmen Muñoz González said Operation Shams II showed the value of international law-enforcement cooperation combined with responsible AI-assisted analysis.

The organization has not said which AI agents, models, vendors, or hosting locations were involved. “AI-generated scripts” could refer to automation code created with help from a large language model rather than a purpose-built system. A related Interpol project, DTECH-Light, was reported in 2019 as a system intended to detect, extract, and analyze digital terrorist content.