Open Journal Systems

Calls for Criminal Justice Reform

Among the many issues that were at stake in last week’s election, it was a chance for voters to weigh in on a contentious national debate over the criminal justice system. The killing of George Floyd in May marked the beginning of sustained protests over deaths of Black Americans at the hands of law enforcement. Both major party presidential candidates claimed they would reform the criminal justice system.[1] Setting aside the credibility of these claims, they represented an acknowledgement that most Americans are unsatisfied with the status quo.

But the political right and the political left have starkly different goals in mind for the criminal justice system. The difference in rhetoric is obvious, as evidenced by simple slogans such as “law and order” and “defund the police.” Underlying these slogans is a more substantive debate. Some Americans want to reinforce traditional approaches to policing and prosecution, while others want to pursue progressive reforms that prioritize reducing incarceration, treating defendants therapeutically, and addressing social problems that lead to crime.

For those who support the progressive approach, the election of Joe Biden as President is a crucial victory. Another victory, which has not grabbed very many headlines, is that many progressive prosecutors were elected in down-ballot races across the country.[2] This new cohort of progressive prosecutors will have a profound impact on their communities, as prosecutors exercise significant discretion over how to charge defendants for crimes and when to pursue alternatives to incarceration. Indeed, some academic commentators have argued that prosecutors are an even greater driving force in the criminal justice system than judges and legislators.[3]

To temper the expectations of progressives, however, it is worth noting that prosecutors elected on progressive platforms often lose some of their progressive bona fides once they are in office. For example, there has been a great deal of debate over whether Kamala Harris was a progressive prosecutor, notwithstanding that she was elected as District Attorney of San Francisco on a very progressive platform—replacing a predecessor, by the way, who already saw himself as progressive.[4] The District Attorney of New York County (Manhattan), Cyrus Vance, Jr., is a former public defender who was also elected on a progressive platform. Yet the candidates vying to replace him in an upcoming 2021 election are styling themselves as more progressive than he is.[5]

Progressive prosecution is an important concept, but not a very specific one. It plays a role in how prosecutors get elected and it may be useful for describing a culture that is becoming more common in prosecutor’s offices. Once prosecutors are in office, however, they are often faced with difficult choices that do not fit neatly within the paradigms of traditional prosecution or progressive prosecution.

Intelligence-Driven Prosecution

One of the choices facing the country’s newly-elected progressive prosecutors is whether to adopt the framework of intelligence-driven prosecution. Intelligence-driven prosecution started a decade ago in the New York County District Attorney’s Office and it is quickly spreading across the country.[6] It represents a new way of thinking about the role of the prosecutor.

The traditional prosecutor brings charges against a defendant based only on whether they committed the elements of the crime. An intelligence-driven prosecutor also asks who the defendant is and what their impact on public safety is, and tailors their charges accordingly. Suppose two people commit similar crimes of larceny; one has never committed a crime before, while the other has committed two violent crimes in the past year. In a traditional prosecutor’s office, they may receive the same charges. In an intelligence-driven prosecutor’s office, the person offending for the first time is more likely to be treated therapeutically; the person who has committed two recent violent crimes is more likely to be deemed a risk to public safety, given more elevated charges, and ultimately incarcerated.

To progressive prosecutors, intelligence-driven prosecution presents a difficult trade-off. On one hand, it can help make progressive prosecution possible, as it allows prosecutors to pursue alternatives to incarceration while having some confidence that the individuals they treat therapeutically are not the ones who are most responsible for driving violent crime in their communities. Meanwhile, the small number of individuals who are driving violent crime can be given appropriately harsh treatment. Thus, intelligence-driven prosecution can provide a targeted approach that, if executed effectively, can reduce incarceration and improve public safety at the same time. Ultimately, that kind of success also makes it more likely that progressive prosecutors will stay in office and progressive policies will persist.

On the other hand, intelligence-driven prosecution raises concerns over bias. In order to know how individuals are connected to crime, an intelligence-driven prosecutor’s office uses data to track crime hotspots and individuals who are driving crime.[7] The concern that this kind of tracking may introduce bias is analogous to concerns over bias in the use of predictive algorithms in policing. Critics have argued that predictive algorithms are likely to train themselves based on communities that are more heavily policed, and that this may have the effect of reinforcing historical patterns of policing.[8] This past summer, Santa Cruz became the first city to ban predictive policing altogether, out of concern that such technologies can contribute to racial inequalities.[9]

Possible Solutions

Many prosecutors have chosen not to pursue intelligence-driven prosecution at all, sometimes out of concern over bias. These prosecutors continue to make charging decisions based only on whether the elements of the crime were committed.

Other prosecutors attempt to find a middle ground, pursuing intelligence-driven prosecution while also minimizing the collection of data within the prosecutor’s office. This approach removes concerns over bias resulting from data, but it is also likely to undercut some of the benefits of intelligence-driven prosecution. After all, if charging decisions are based only on the information collected for an individual case, prosecutors may not know which individuals are driving crime trends, and they may therefore not be able to make charging decisions in a targeted fashion. It is also important to note that this approach does not remove concerns over human bias.

Many prosecutors are embracing intelligence-driven prosecution, and they must continue to find ways to minimize the risk of bias. One way to minimize the risk of bias may be to avoid the use of algorithms or formulas in making prosecutorial decisions, and instead to simply make data available to prosecutors when deciding how to charge a defendant. Under this approach, data is only used to provide context to a human decision-maker. While human bias is still possible, perhaps data serves as a check on human bias when used in this fashion. Another way to minimize the risk of bias may be to create stringent standards for the quality of data that can inform prosecutorial decisions. After all, the prosecutor’s decision as to what charges to bring against a defendant can have as much of an effect on the defendant’s future as a guilty verdict at trial. While there are strict rules governing the quality and nature of evidence that may be admitted at trial, the rules for data that informs a prosecutor’s charging decision are not as clear. A prosecutor’s office can define an appropriately high standard of reliability and hold its data to that standard.

Given that progressive prosecution and intelligence-driven prosecution are both on the rise, these tensions will continue to play out over the coming years. It remains to be seen how significant the risk of bias in intelligence-driven prosecution is, and to what extent this risk can be minimized. If there is an approach to intelligence-driven prosecution that allows prosecutors to make targeted decisions without introducing concerns over bias, that may be the holy grail for progressive prosecutors. For now, we are in a period of experimentation.




[3] See, most notably, William J. Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505 (2001).