Policing by camera and algorithm
Policing, law enforcement and crime appear on 63 recent papers, with 18 in the last quarter against about 15 per quarter before, but only 12 carry a technology, surveillance, camera or algorithm keyword, and "predictive policing" on 3. Cameras, licence plate readers and risk scores are spreading fast. Do they cut crime, and at what cost to whom?
What prompted this
Since October 2025, 63 papers carry a police, policing, law-enforcement, crime-prevention or criminal-justice keyword (33 in economics journals, 16 in AI): "law enforcement" on 13, "crime rates" on 8, "crime prevention" on 7, "policing" on 5, "predictive policing" on 3. Only 12 of the 63 also carry a technology, surveillance, camera, algorithm or AI keyword.
Questions
- Do licence plate readers and camera networks reduce crime, displace it, or mainly raise clearance rates?
- Do algorithmic risk scores change who is stopped and charged, and do they reduce or reproduce disparities?
- Did body cameras change complaints and use of force in the long run, once the novelty wore off?
Data that could help
Crime and arrest records at the beat level, dates of technology deployments by police department, and complaint records.
Next step
Use the staggered installation of licence plate reader networks across US cities to estimate changes in vehicle theft and its displacement to neighbouring areas.