Does AI Generate a Civilizing or De-policing Effect? Testing the Impact of AI-Based Body Worn Camera Analytics in a Large State Police Agency

AI-based body-worn camera (BWC) analytics has the potential to impact police officer behavior by dramatically expanding footage review. The influence on officers could be positive (better behavior through enhanced oversight and automated feedback; a civilizing effect) or negative (reduced activity from over-surveillance; de-policing). However, the evidence on the technology is limited. The present study employs a randomized controlled trial to test the impact of Truleo (an AI-based BWC analytics platform) on a range of performance measures among troopers in a large state police agency. We use count-based regression models to assess the treatment effect on activity measures (stops, citations, arrests, use of force, complaints) and mixed effects regression models for a Truleo-generated professionalism measure. We also assess alignment between the activity and professionalism measures. We find no statistically significant effects on activity measures after multiple-comparison correction. There was a modest nonsignificant increase in High Professionalism among the Treatment group. Stops and citations were associated with High Professionalism, though arrests, complaints, and use of force were not. AI may produce important organizational efficiencies in policing, but the findings here suggest access to Truleo’s AI platform had limited impact on trooper field behavior (no evidence of a civilizing or de-policing effect). More research is needed in this area given the novelty and proliferation of such platforms.