Answers are grounded in this paper and cite their sources.
This systematic review evaluates research on artificial intelligence (AI) crime-fighting applications used in U.S. policing. Using the “What Works, What Doesn’t, What’s Promising” framework and the Maryland Scientific Methods Scale (SMS) to assess evidence strength, the review synthesizes findings from 19 empirical studies. Results show that place-based predictive policing can reduce crime, harmspot policing may lower social harm without increasing disparities, and facial-recognition systems can improve investigations and case clearance when supported by strong accountability. In contrast, person-based targeting, license plate readers, acoustic gunshot detection, and AIassisted crime-report writing show limited or mixed effectiveness. Several emerging AI technologies appear operationally promising but lack outcome-based evaluations needed to determine their real impact on crime reduction. The review highlights the importance of evidence-based AI implementation in law enforcement and calls for more rigorous research to assess both the effectiveness and risks of AI-driven crime-fighting tools.
No takes yet. Share an insight, caveat, or question.
Boke et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: