INSPECTr, GRACE, and COPKIT all focus on extracting, correlating, and analyzing digital evidence for criminal investigations.
POLICE FEDERALE BELGE
Belgium's federal police force contributing operational law enforcement expertise to EU security research in AI, digital forensics, and cybercrime.
Their core work
The Belgian Federal Police is Belgium's national law enforcement agency, contributing operational expertise and real-world policing requirements to EU security research. In H2020 projects, they serve as an end-user validating tools for digital forensics, counter-terrorism intelligence, cybercrime investigation, and border security document fraud detection. Their involvement ensures that research outputs — from AI-driven evidence analysis to autonomous reconnaissance systems — meet the practical needs of frontline law enforcement across Europe.
What they specialise in
COPKIT addresses early-warning policing against organized crime and terrorism, while STARLIGHT tackles high-priority threats using AI.
CYCLOPES builds a cybercrime practitioners' network, and STARLIGHT focuses on technological autonomy and cybersecurity for law enforcement.
iMARS develops solutions for detecting image manipulation and morphing attacks on identity documents.
INTREPID — their largest-funded project — explores autonomous robotics, extended reality, and situational awareness for perilous incident response.
GRACE applies NLP, computer vision, and federated learning to combat child sexual exploitation material.
How they've shifted over time
Early projects (2018–2019) focused on intelligence-led policing: knowledge discovery, OSINT, spatial-temporal prediction, and counter-terrorism early warning systems. From 2020 onward, the focus shifted decisively toward operational technology: autonomous robotics for dangerous situations, AI for detecting manipulated identity documents, and AI-powered tools against cybercrime and online child exploitation. This trajectory shows a move from analytical intelligence tools toward field-deployable AI systems and specialized crime-fighting technologies.
Moving toward AI and autonomous systems for frontline policing — expect growing interest in responsible AI, real-time decision support, and ethical AI governance for law enforcement.
How they like to work
The Belgian Federal Police participates exclusively as a partner, never as coordinator — consistent with their role as an end-user providing operational requirements and validation rather than managing research. They work in large consortia (129 unique partners across 7 projects, averaging ~18 partners per project), which reflects the scale of EU security research calls. Their broad partner network suggests they are a sought-after end-user voice rather than a repeat collaborator with a fixed set of research partners.
Extensive network of 129 unique partners across 27 countries, spanning nearly all EU member states. This breadth reflects their role as a valued law enforcement end-user in large, pan-European security consortia.
What sets them apart
As a national police force, they bring something academic partners cannot: real operational context, access to authentic use cases, and authority to validate whether a tool actually works in the field. Their participation signals to the European Commission that a project has genuine end-user commitment, not just theoretical relevance. For consortium builders, having them on board strengthens both the proposal's credibility and its path to real-world deployment.
Highlights from their portfolio
- INTREPIDTheir largest-funded project (EUR 264,450) combining autonomous robotics, extended reality, and tactical reconnaissance — an unusually hands-on operational technology scope.
- GRACEApplies federated learning and computer vision to combat child exploitation — a sensitive domain where law enforcement end-user input is critical for ethical and effective tool design.
- STARLIGHTA flagship project on AI sovereignty and resilience for European law enforcement, addressing the strategic question of technological autonomy in policing.