AI Politie
© Lotte Debrauwer

AMSTERDAM / GHENT – Can a burglary be predicted before it happens? According to the ambitious Bigdatpol project at Ghent University, this will soon become a reality. Led by Professor Wim Hardyns, the research team has spent years developing an AI model that can identify areas and times of high burglary risk.

The aim is to deploy patrols to these areas in a strategic manner in order to reduce crime through a deterrent effect. Trials are planned for 2026 in at least twenty Flemish police districts.

In this project, journalist Lotte Debrauwer delves deeper into the subject. Through open-source research, interviews with experts, and by acting as a police officer herself, she exposes the risks associated with this predictive policing technology. She demonstrates why the model is not neutral and explains why the trials in Flanders may take a long time to materialise.

She also turns her attention to the Netherlands, where the Dutch police phased out the Crime Anticipation System (CAS) after more than a decade. An in-depth investigation based on freedom of information requests reveals just how problematic the use of the CAS model was for Belgium's northern neighbours.

Supported
€4,500 allocated on 08/12/2025
ID:
FPD/2025/2450

Team members

Need resources for your own investigative story?

Journalismfund Europe's flexible grants programmes enable journalists to produce relevant public interest stories with a European mind-set from international, national, and regional perspectives.

Support independent cross-border investigative journalism

We rely on your support to continue the work that we do. Make a gift of any amount today.