New publication about the influence of (explainable) AI on police officers’ judgment

Do recommendations generated by AI (such as predictive policing systems) alter the judgment of police officers? In a recently published scientific article Friso Selten (Leiden University), Marcel Robeer (Utrecht University & National Police Lab AI) and Stephan Grimmelikhuijsen (Utrecht University) find that police officers still follow their intuitive professional judgement even when a predictive AI system prompts them otherwise. The AI system is only trusted when it confirms prior judgements, even when a recommendation is supported by explainable AI (XAI).

More information about this study, research design and conclusions can be found here