Washington Law Review


In the last five years, legislators in all fifty states have made changes to their pretrial justice systems. Reform efforts aim to shrink jails by incarcerating fewer people—particularly poor, low-risk defendants and racial minorities. Many jurisdictions are embracing pretrial risk assessment instruments—statistical tools that use historical data to forecast which defendants can safely be released—as a centerpiece of reform. Now, many are questioning the extent to which pretrial risk assessment instruments actually serve reform goals. Existing scholarship and debate centers on how the instruments themselves may reinforce racial disparities and on how their opaque algorithms may frustrate due process interests. This Article highlights three underlying challenges that have yet to receive the attention they require. First, today’s risk assessment tools lead to what we term “zombie predictions.” That is, predictive models trained on data from older bail regimes are blind to the risk-reducing benefits of recent bail reforms. This may cause predictions that systematically overestimate risk. Second, “decision-making frameworks” that mediate the court system’s use of risk estimates embody crucial moral judgments, yet currently escape appropriate public scrutiny. Third, in the long-term, these tools risk giving an imprimatur of scientific objectivity to ill-defined concepts of “dangerousness,” may entrench the Supreme Court’s historically recent blessing of preventive detention for dangerousness, and could pave the way for an increase in preventive detention. Pretrial risk assessment instruments, as they are currently built and used, cannot safely be assumed to support reformist goals of reducing incarceration and addressing racial and poverty-based inequities. This Article contends that system stakeholders who share those goals are best off focusing their reformist energies on other steps that can more directly promote decarceral changes and greater equity in pretrial justice. Where pretrial risk assessments remain in use, this Article proposes two vital steps that should be seen as minimally necessary to address the challenges surfaced. First, where they choose to embrace risk assessment, jurisdictions must carefully define what they wish to predict, gather and use local, recent data, and continuously update and calibrate any model on which they choose to rely, investing in a robust data infrastructure where necessary to meet these goals. Second, instruments and frameworks must be subject to strong, inclusive governance.

First Page