(Re)Individualizing Criminal Law

Criminal law purports to care about the individual culpability of those committing crimes. Individual mental states or defenses based on one’s beliefs are considered foundational to moral responsibility and criminal conviction. The expansion of algorithmic risk assessment tools in the criminal legal system, however, threatens these principles, changing the criminal law subject from an autonomous individual to merely a predicted object. This Article argues that risk assessment tools—now functioning as gatekeepers at both the entry and exit stages of the criminal process, from predictive policing to compassionate release eligibility—have shifted the criminal system away from its foundation in individual culpability.

     Scholarship on risk assessment tools has thus far largely analyzed their isolated impact on specific stages of the criminal process while portraying algorithmic tools as being bound by existing laws. This Article bridges these siloed discussions and demonstrates that far from being bound by existing laws, algorithmic tools threaten fundamental criminal law principles and reshape substantive criminal law itself. It thus uncovers the aggregated, overlooked harms these tools inflict on the foundations of modern criminal law. The Article then offers a novel legal framework to tackle these harms. By integrating the right to due process, the right to contest AI, and the right to speak in criminal proceedings, it proposes a system of “re-individualization through contestation.” This framework will not only recalibrate algorithmic tools to accord with fundamental criminal law principles, but also offer an opportunity to leverage mounting criticism of unchecked algorithmic expansion to catalyze solutions to decades-old problems in our criminal legal system.