EPIC to Mass. Supreme Court: Block Use of Opaque & Unfair Parole Risk Assessments

EPIC has filed an amicus brief in Rodriguez v. Massachusetts Parole Board arguing that the risk assessment tool the Board relies on to make parole decisions is unfair, opaque, and inappropriate for use on juvenile offenders. The plaintiff in the case, Jose Rodriguez, is a juvenile lifer seeking to reverse a denial of parole that was based on in part on the LS/CMI, a risk assessment tool relied on by the Massachusetts Parole Board. Risk assessment tools are automated decision-making systems used by most states that purport to predict the likelihood of certain outcomes, such as the rearrest of a parolee. But the LS/CMI has low measures of predictive accuracy even across the general population of incarcerated people, let alone smaller subgroups. When applied to juvenile lifers, whose life experiences and circumstances differ drastically from the average incarcerated person, there is reason to believe that LS/CMI’s predictions are even less accurate. Massachusetts does not require localized and independent testing of the accuracy or bias of the tool, and both the Parole Board and the company that developed the LS/CMI have withheld vital information about the tool from parolees and the public. EPIC regularly advocates for algorithmic transparency and accountability. In 2020, EPIC published Liberty At Risk, a report with public records, a state-by-state matrix of risk assessment tools, and policy recommendations.