UK Police Predictive Crime Tool Produced Untrustworthy Results
June 25, 2026
A WIRED investigation into a UK regional police force's predictive analytics deployment found reliability problems with outputs, raising questions about operational use of crime-prediction AI. The findings highlight gaps between vendor claims and real-world performance in high-stakes public safety contexts.
HOW THIS AFFECTS YOU
●
policyThis is a concrete documented case of predictive policing AI failing reliability checks in production, useful evidence for AI governance and procurement standards debates.