The Username of the Father: Predictive Sentencing Algorithms, Race, and the Shame of Law

• In Event: Algorithms at Work: The Practice of Prediction - III: Reframing Algorithmic Accountability in Practice

Abstract

Over the past fifteen years, predictive algorithms have become increasingly common in American courts. These proprietary software programs use data about defendants and newly convicted criminals to calculate their likelihood of recidivism and future violence, which in turn informs judicial decision-making about bail, parole, and sentencing. In this paper I call attention to COMPAS, the most well-known such algorithm whose opacity and evident racial bias in high-profile cases has provoked an outcry from academics, journalists, and the general public. I first frame this concerned discourse as emblematic of the present moment, in which awareness of racial disparity and attentiveness to the dangers of technology increasingly define mainstream thought in popular culture. While acknowledging the validity and sincerity of this mode of critique, I then consider what work this rhetoric performs in protecting both the state and the public from confronting its brutal system of racial domination– what I term the shame of law. Asserting that the essence of law is not reason but force itself, I argue the judiciary provides a veneer of proceduralism that serves as a buffer protecting the public from confronting this dark reality. Algorithms like COMPAS, I conclude, weaken this buffer by extending proceduralism to its absurd conclusion– creating a caricature of the judge as calculating, deliberate, and rational and in so doing threatening to undo this intricate legal fiction.

Author

• Aaron Neiman, Stanford University