Last August saw the historic events of the Michael Brown shooting unfold in dramatic fashion on the streets of Ferguson and in headlines across the country. They brought up a very old conversation, police race relations, with a renewed sense of urgency. At its heart, this issue is a crisis of confidence in the trustworthiness of the police—a major problem for an institution that exists to serve the people and relies on their support.

In the search for solutions, technology has been a popular option. Dashboard or body cameras have been tried to mixed success in police departments across the country, but issues abound. In some cases, the cameras can simply be turned off. Even when the cameras do capture alleged misconduct, questions remain about how useful that video could even be. That fact was thrown into sharp relief by a New York jury’s decision not to indict the officer involved in the death of Eric Garner, despite the existence of video documenting the incident. I propose that the root problem is one of incentive—simply put, for all their good and noble intentions, police have a strong disincentive to document their own wrongdoing.

Meanwhile, the party with the strongest incentive to make progress on this problem is the community itself, especially its marginalized segments. Thus, I propose a new technology which would empower these groups to act in their own interest: CopWatch.

CopWatch would be a fairly simple mobile app. Users could use the app to report an interaction with the police. GPS data would be taken from the phone, and the user could specify the nature of the interaction (traffic stop, stop and frisk, etc…) and a “satisfaction rating.” The app would also support mobile video or photo uploads to document the interaction.

This data could be used to map not just police action, but also the nature of those actions and the effects that those actions have on the local community. It could be the case, for example, that two neighborhoods might have similar levels of police interaction, but while one neighborhood saw almost exclusively positive interactions, the other was dominated by negative ones. These maps could be compared against crime and demographic data to see where police are over patrolling relative to the population and crime rate.

By itself, data doesn’t do much. But trends uncovered by the data could help inform policy decisions to help tackle these problems. Data could also help justify police action to citizens when it is legitimately warranted; after all, the burden of police-community relations does not fall only on the police. This data could help police prove the effectiveness of police tactics, which would also help bolster public trust.

A dataset like that generated by CopWatch could also help a nation capitalize on differences between existing policing regimes. Perhaps the police in Denver have certain very effective policies in place, while police in Cleveland are not doing as well. CopWatch data could help identify model districts and facilitate the adoption of good police techniques. This would help turn the diversity of policing styles inherent in a large nation like America into a major asset.

It is unlikely that any single solution could hope to tackle such a huge problem, but it is almost certain that police and citizens alike will be turning to technology for help. Technology that empowers the groups with the strongest incentive to improve the status quo will likely prove to be a powerful tool, be it in the form of CopWatch or something else. If thoughtfully constructed, such technology also stands to improve policing altogether.