Algorithmic Authority: The Perils of Computational Selection

This paper analyses and discusses the entrenched place automated, algorithmic decision-making has come to occupy in the authoritative arsenal of state immigration control. It does so in service of a wider normative point about the legitimacy of authoritative statist migration control as such, which I understand to be increasingly undermined the more key control policies and practices fail to recognize and respect their subjects' inherent autonomy and moral equality. I point to several ways in which 'algorithmic authority' strains the inherent autonomy and moral equality of migration control subjects, arguing that these also compound each other. In particular, I will shed light on four characteristics of algorithmic authority that come together to undermine either would-be immigrants’ autonomy or their moral equality, or both: comprehensive surveillance, heteronomous identity construction, discrimination exacerbation, and opacity. While these are all relevant and interrelated, my purposes will be served by a particular focus on the latter two. Ultimately, there is a plethora of normative reasons to advocate the reform of algorithmic decision-making processes, and increasing doubt that migration control systems which rely heavily on algorithmic authority as currently practiced can claim legitimate authority over (would-be) immigrants are fully justified. These conclusions, in turn, have implications for the moral evaluation of migrant subversion and resistance.