"Progressive Algorithms"

The title of this post is the title of this notable new paper authored by Itay Ravid and Amit Haim available via SSRN.  Here is its abstract:

Our criminal justice system is broken.  Problems of mass incarceration, racial disparities, and susceptibility to error are prevalent in all phases of the criminal process.  Recently, two dominant trends that aspire to tackle these fundamental problems have emerged in the criminal justice system: progressive prosecution — often defined as elected reform-minded prosecutors that advance systemic change in criminal justice — and algorithmic decision-making — characterized by the adoption of statistical modeling and computational methodology to predict outcomes in criminal contexts.

While there are growing bodies of literature on each of these two trends, thus far they have not been discussed in tandem.  This Article argues that scholarship on criminal justice reform must consider both developments and strive to reconcile them.  We argue that while both trends promise to address similar key flaws in the criminal justice system, they send diametrically opposed messages with respect to the role of humans in advancing criminal justice reform.  Progressive prosecution posits that humans are the solution, while algorithmic tools suggest humans are the problem.  This clash reflects both normative frictions and deep differences in the modus operandi of each of these paradigms.  Such tensions are not only theoretical but have immediate practical implications such that each approach tends to inhibit the advantages of the other with respect to bettering the criminal justice system.

We argue against disjointly embedding progressive agendas and algorithmic tools in criminal justice systems.  Instead, we offer a decision-making model which prioritizes principles of accountability, transparency, and democratization without neglecting the benefits of computational methods and technology.  Overall, this article offers a framework to start thinking through the inherent frictions between progressive prosecution and algorithmic decision-making and the potential ways to overcome them.  More broadly, the Article contributes to the discussions about the role of humans in advancing legal reforms in an era of pervading technology.