Skip to main content

How Automation in Public Safety Operations Can Help Reduce Bias

As we rethink transparency and accountability in policing, more agencies are recognizing that they can effectively reduce bias by accelerating the adoption of automated traffic violation enforcement.

Currently, automated speed enforcement systems are used by more than 90 jurisdictions. In many cases, agencies explicitly prohibit the use of facial recognition, and in most programs they do not capture the driver or passenger faces, genders, or ethnicities of people in the car. This fact helps alleviate concerns about privacy as well as bias. Indeed, the vast majority of state-enabling legislation for camera-based enforcement does not require driver identification.  

The need for police and public interaction roadside is not necessary with automated enforcement systems that are installed at fixed locations or in mobile units. The systems use sophisticated data and imaging methods to capture violations based on vehicle speed alone. This is critically important to reducing escalation risks inherent in interactions between police and the public. In recent years, the number of such interactions have declined. However, traffic stops remain the single largest reason people interact with police officers. And the single largest reason for traffic stops is speeding. Rather than casting a wide net, which is often perceived as targeting specific groups, we recommend to our clients that police focus on known egregious criminals rather than manually profiling people during traffic stops, and deploy automated traffic enforcement systems.  

We are all human and no matter how hard we work to avoid bias, humans are judgmental and more error prone than automated systems that track vehicle velocity. Research on enforcement confirms this theory, finding that bias is evident when human-issued tickets are compared with those captured by cameras.

We suggest agencies use impartial and well-documented methods when determining where to locate traffic cameras. At a minimum, agencies should use detailed traffic information and proximity to at-risk populations in determining traffic camera placement. Agencies can also use data such as intersections with the highest percentage of serious crashes, or fatalities relative to traffic volume.  Northwestern University produced an independent study that outlined a procedure for location determination for red light cameras. Such independent studies provide guidelines that can help municipalities ensure a reasonable and problem-focused program. Another practical enforcement approach that many municipalities deploy is enforcing the speed limit near locations where people are most at risk, such as road work zones, school zones, and parks.  Adding speed enforcement in these areas reduces risks to vulnerable children and workers who must function in dangerous locations.

Using automated processes and sensible data-driven deployment of camera systems will reduce bias in policing, while also helping at-risk populations, enabling police officers to focus their limited resources on distinct community threats.