New APPS

Forthcoming in Philosophy and Social Criticism; preprint here on SSRN. Here is the abstract:

“Algorithmic governance is sometimes compared to Hobbes’s Leviathan. Here I argue that, while algorithmic governance shares some similarities with the Hobbesian schema, it goes further in its suppression of contestation. To show this, I read Hobbes against Rancière’s reduction of him to a theorist of consensus to make two basic points. First, it is important for Hobbes that laws be public and rationally comprehensible. By contrast, algorithmic governance is notoriously opaque. Second, Hobbes also retains a sense of equity as an appeal against universally applied legal decisions, allowing decisions to be tailored to individual cases. Algorithmic governance does not typically involve equity-based appeal, a point that is especially clear in the context of bureaucratic governance structures, where algorithmic systems generate results that cannot be either understood or appealed. The result in Rancière’s terms is that classificatory, algorithmic systems are even more powerful agents of depoliticization than Leviathan.”

And here’s a little more detail:

“Part 2 develops an account of equity as a historic principle of law.  Part 3 is about Hobbes and focuses on the extent to which he emphasizes the public nature of law and retains an Aristotelian view of equity.  Part 4 turns to algorithmic governance.  I begin with fundamental epistemic differences between Hobbesian and algorithmic thought and outline three fundamental implications of them.  First, the sense in which algorithmic governance is inescapable is different from that in which the Leviathan is inescapable because it is much harder to contest algorithmic decisions.  Second, algorithms are less public because of their opacity.  Third, algorithmic efficiency undermines equitable implementation of policy by shifting the decisionmaking of street-level bureaucrats toward algorithms and away from the exercise of human judgment.  Collectively, these show how the integration of algorithmic processes in government effectively erodes the rule of law features that were prominent even in Hobbes.  Section 5 offers a brief conclusion.  I return to Rancière’s contention that Hobbesian governance produces an outside of those who are equal insofar as they have no part in the Leviathan-state.  Insofar as algorithmic governance is classificatory, and insofar as its classificatory regimes proliferate indefinitely, even for the same person, the classificatory process works against this final vestige of equality and political contestation”

(the parts are named rather than numbered in the final version; this is just a preprint)