With police reform at the forefront of conversations at the federal, state, and local levels, there is growing recognition that improving policing in the United States requires rethinking the role of police in traffic stops. Currently, traffic stops are the most common way that police come into contact with civilians in the United States. Traffic stops are also a major setting in which civilians, and especially people of color, are vulnerable to being subjected to police violence. The tragic killings of Daunte Wright, Philando Castile, and many others killed by police during traffic stops illustrate that stops for minor traffic violations can quickly escalate into deadly encounters for unarmed people of color who pose no threat.
In the United States, the responsibility to enforce traffic laws is largely placed on armed law enforcement officers with traditional police powers to detain, search, and arrest. Several cities and localities, however, are considering alternative approaches to traffic enforcement that do not rely on police. Most notably, in July 2020, the city of Berkeley, California voted in favor of a groundbreaking proposal that directs the city to create a department of transportation staffed by unarmed civil servants who would be in charge of enforcing traffic laws instead of police officers.
As calls to remove police from traffic enforcement grow, there is increasing interest and curiosity in whether technology can have a positive role in supporting traffic policing reform, either in the near-term or the long-term. In a forthcoming article, “Traffic Without the Police,” soon to be published in the Stanford Law Review, I outline a framework for removing traffic enforcement from the police that, with caution, leaves room for cities and localities to embrace automation as a partial solution to reduce traffic stops conducted by police. Research from other scholars also stresses the potential benefits of using automation to replace police-initiated traffic stops.
Hundreds of cities and localities have already established red-light or speed camera enforcement programs. However, many communities are retreating from those programs in response to community and political backlash. Automated traffic enforcement programs have also been subject to legal challenges in courts with varying success. The strong skepticism of and pushback against automated traffic enforcement raises questions about whether and how technology can assist in reducing reliance on police-initiated traffic stops to enforce traffic laws.
A key benefit of automated traffic enforcement is that it eliminates the potential for traffic stops based on minor traffic violations to escalate into unjustified and unnecessary police violence, especially against people of color. Ultimately, however, there are important reasons to be skeptical about whether automated traffic enforcement is the best method of achieving this benefit over other non-police alternatives to traffic enforcement, such as creating new civil traffic enforcement agencies. In the discussion below, I explain why this is the case and outline seven key challenges that state and local governments must consider for automated traffic enforcement to succeed as a method of police reform.
First, state and local governments must be sensitive and responsive to the fact that their constituencies might view automated traffic enforcement as a threat to privacy and an additional tool of unwanted government surveillance and control. Automated traffic enforcement poses serious privacy concerns depending on how information is collected, stored, and used. One major concern is that state and local governments will gather and store data involving a driver’s whereabouts, regardless of whether a citation is issued or successfully appealed in a given case. Government and private entities could then potentially use that information for purposes other than issuing automated traffic citations (for instance, criminal prosecutions, criminal investigations, or civil lawsuits).
These issues are especially important to consider for communities of color and other marginalized communities. Historically, governmental entities have used new surveillance technologies to disproportionately target and monitor these groups.In the context of policing, this history is increasingly relevant the more we learn about automation and racial bias. For instance, recent research reveals that facial recognition systems are more likely to inaccurately identify people of color and can exhibit similar racial biases held by humans.
Given these concerns, government and private actors should not have access or be able to use information collected through automated traffic enforcement for purposes other than issuing automated traffic citations and the subsequent appeals process. Moreover, information should be destroyed as soon as possible after automated traffic enforcement cases are resolved. Laws and policies should require these practices.
Second, state and local governments must remove police agencies and police officers from the management and control of automated traffic enforcement programs. Currently, automated traffic enforcement is a secondary tool of the police. This approach raises concerns over what police agencies can do with the information that they receive from traffic cameras and who has access to that information. These concerns are especially salient given that police today use traffic enforcement as a tool of criminal investigation, and often in pretextual ways that target communities of color. In my research, I argue that to the extent that jurisdictions automate certain aspects of traffic enforcement, newly-created traffic agencies that operate independently from the police should manage and handle all aspects of automated traffic enforcement programs.
Third, traffic safety must guide decisions about where automated traffic enforcement cameras are placed in particular communities. These decisions must also be made in transparent ways. As a means of achieving traffic safety, traffic cameras should only be placed at intersections or on roads that data indicate are the highest risk for traffic accidents or crashes. Moreover, traffic cameras should only target the specific driving behaviors that lead to accidents or crashes in those spaces (such as blowing through red-lights or excessive speeding). This requires that traffic cameras be equipped to accurately detect those specifically risky driving behaviors as opposed to mere technical violations of traffic laws. For instance, traffic cameras should not issue tickets as a matter of course when a car merely enters an intersection on a red light or is driving only a few miles an hour above the speed limit without any imminent public safety risk.
Fourth, states and localities must take affirmative steps to ensure that the placement of traffic cameras does not disproportionately burden low-income and communities of color. Recent research from the District of Columbia reveals how communities of color can bear the brunt of automated traffic enforcement programs when traffic cameras are disproportionately placed in neighborhoods with high concentrations of people of color. If data shows that the intersections and roads at greatest risk for traffic accidents and collisions are in neighborhoods with higher concentrations of people of color, then states and localities should explore whether changes in road design (for instance, new signage or road construction) could improve traffic safety as an alternative solution.
Fifth, automated traffic enforcement programs should not support profit motives or serve as revenue-generating tools for state and local governments. States and localities benefit from aggressive and biased traffic enforcement by using traffic ticket revenue to fund their respective budgets. These aggressive and biased ticketing practices often fall hardest on low-income people and communities of color. Automated traffic enforcement could exacerbate these problems by dramatically increasing the volume of issued traffic tickets with little to no human involvement.
To address these concerns, states and localities must create systems of oversight to monitor the extent to which automation changes the volume of issued traffic tickets and the extent to which those changes affect low-income people and communities of color. Limiting where traffic cameras are placed and which traffic violations those cameras capture can help to limit the volume of automated traffic tickets. To the extent that state and local governments rely on third-party vendors to operate automated traffic enforcement programs, compensation should not be based on the volume of automated traffic tickets. Ideally, newly created traffic agencies would hire and train their own staff to process automated traffic tickets rather than relying on third-party vendors.
Sixth, automated traffic enforcement must not perpetuate the criminalization of poverty. When drivers violate traffic laws, they typically face the same financial penalties regardless of their income or ability to pay. For many people, the cost of a simple traffic ticket is beyond their means. Unpaid traffic debt or failure to appear in court for a traffic ticket can result in heavy additional financial penalties, loss of a driver’s license, garnished wages, bench warrants for arrest, and even incarceration in many states. States and localities must take steps to eliminate possibilities for drivers to be funneled into the criminal justice system through unpaid automated traffic enforcement debt. Potential steps include requiring consideration of a driver’s ability to pay in assessing financial penalties for automated traffic tickets and ending revocation of driver’s licenses, bench warrants, and the possibility of incarceration for unpaid automated traffic debt.
Seventh, automated traffic enforcement programs must give prompt notice to drivers of any issued traffic tickets and their underlying infractions and have meaningful and accessible appeal processes for drivers to challenge those tickets. If it takes weeks for drivers to receive an automated traffic ticket, then drivers may forget the circumstances surrounding the alleged traffic infractions, making it more difficult for them to challenge their tickets on appeal. Traffic agencies should dispose of any automated traffic infractions that cannot be promptly processed and handled.
Given these challenges, automation may not be the best solution to improve traffic enforcement in all communities. Ultimately, there are legitimate reasons why communities may prefer to pursue other non-police alternatives either in the near-term or the long-term to stop traffic enforcement from escalating into unjustified and unnecessary violence against people of color. However, the recommendations above are useful first steps for state and local governments to increase the potential for automated traffic enforcement to be a successful measure of police reform.