Lau on Algorithmic Authority in Criminal Justice

Danny Lau (Hughes Hall, University of Cambridge; St. Antony’s College, University of Oxford) has posted Legitimacy Under Strain: What the COMPAS Debate Reveals About Algorithmic Authority in Criminal Justice on SSRN. Here is the abstract:

This paper argues that the public dispute over COMPAS, the proprietary recidivism risk-scoring tool used across several American jurisdictions, exposes a deeper tension between two competing logics of institutional legitimacy. The first, familiar from rule-of-law traditions, grounds authority in publicly examinable and challengeable reasons. The second, here termed ‘rule of model’, treats conformity with performance metrics as sufficient warrant for institutional authority, even when the underlying process remains opaque to those subject to it. Drawing on scholarship in science and technology studies and critical data studies, the paper demonstrates that this shift is neither straightforward nor unidirectional: both human judgment and algorithmic systems involve constructed categories, distributed agency, and embedded power relations that resist simple comparison. The paper concludes that the critical variable is not whether the decisionmaker is a judge or an algorithm, but whether the basis of consequential decisions is genuinely open to scrutiny and challenge.

To receive a daily summary of posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.