These Emerging Technologies Will Completely Change The Criminal Justice System

24/05/2017 11:42

What are some of the emerging criminal justice technology innovations that will impact society? originally appeared on Quora - the place to gain and share knowledge, empowering people to learn from others and better understand the world.

Answer by Jennifer Doleac, Founder of the Justice Tech Lab & UVA economics professor:

I think of the potential contributions of technology as falling into three broad categories: reducing crime, improving data quality, and reducing racial inequality. Here are a few innovations that I have my eye on, and why:

1) Applications of machine learning and other types of prediction algorithms. This includes everything from computer-generated risk scores used to determine whether an arrestee should be detained pre-trial, to facial recognition software that identifies someone with an outstanding warrant when they're seen by an officer's body-worn camera. Computers have the ability to crunch data more quickly than the human mind can, and we can more easily control the inputs (for example, consider someone's previous criminal history in the detention decision, but not his race or employment status). This can reduce error resulting from human biases. It won't be perfect, but it could be much better than the status quo. While computer scientists and data scientists think deeply about prediction problems, we now need social scientists to measure the effects of using these tools on outcomes we care about: does it reduce incarceration rates? Recidivism? Racial disparities in arrests or sentencing? The jury is still out on those questions (pun intended).

2) Improved methods of electronic monitoring that can substitute for incarceration. Imagine if instead of sending convicted offenders to prison, we sentenced them to heavy surveillance at home: cameras, audio recording, GPS monitoring, blood-alcohol content monitors, and so on. Depending on their crime (and risk to the community), they might be allowed to leave their home for work or school, or to visit their kids. It's possible that this type of punishment might have less of a deterrent effect than prison does -- if the only penalty for robbing a convenience store is that you'll be forced to sit on your couch and watch Netflix for a few months, maybe more of us would be tempted -- but that's something we could figure out. And there's increasing evidence that incarceration has a criminogenic effect on low-level offenders, so this could be a particularly good option for that population. Mark Kleiman sketched out a version of this high-tech future, emphasizing that surveillance tools allow us to easily scale the amount of supervision to fit an individual's needs, and gradually roll it back to prepare him for release.

3) Sensor technologies improve the quality of information about criminal behavior. One big challenge in studying crime is that we only know about a crime incident if it's reported to police, and we only know if someone reoffends if they get caught. Since reporting and clearance rates are often pretty low, this can be a big problem. I'm constantly on the lookout for technologies that generate more objective and complete data. One example is acoustic gunshot sensors, offered primarily by a firm called ShotSpotter. These sensors detect the sound of gunfire in a particular area, and triangulate the location of the incident (to a precise latitude and longitude) which is immediately sent to the police so they can get to the scene quickly. This could have many benefits, which are so far untested. The one thing it definitely does well is generate a unique dataset containing the full universe of gunfire incidents in a covered area. These data are extremely useful to researchers like myself who want to study the effects of policy interventions on gun violence -- particularly because the vast majority of urban gunfire is not reported via 911. Unfortunately this leads us to a policy problem: in most of the 90 cities where ShotSpotter is implemented, the data generated belong to the firm, not to local residents. This means they are not public record and cannot be shared with the community, journalists, or researchers. (The firm hopes to sell the data, though to whom is unclear.) New data are only valuable if they are available, so in the future local governments need to insist on data ownership in their contracts with tech firms.