Political Equality and Algorithmic Rule

The fundamental problem of political authority is how to reconcile our liberty and moral equality with the fact that some rule over others. In particular, Kolodny (2016) argues that at the core of the problem lies something he calls the ‘subordination complaint’, which is “to be subject to political rule is to be put in relations of subordination, or social inferiority, to other individuals, with whom one has a claim to stand as an equal.” In Kolodnys view, the problem of political authority is a problem of relational (in)equality. What needs explaining is that despite political rulers and political subjects finding themselves in asymmetrical relations, political rule does not hinder per se that they are equals to one another. The fact that some members of a society decide, rule and judge over the other members, who are thereby exposed to the greater power and authority of the former, calls for justification. But there might be a way by which we are able to eliminate the problem. Instead of having other people decide, rule and judge over us, we might now be able to have algorithmic systems decide, rule, and judge over us. Examples are algorithmic systems that decide on appropriate sentencing and whether people qualify for parole, where police patrol and whom they search, whether people qualify for welfare, who gets invited for a job interview, or whether their employment should be discontinued based on their performance on the job. In other words, we can already (or very soon) substitute human rule for algorithmic rule. So can the problem of political authority be solved by automating political rule? Does algorithmic rule make us more equal? On the view introduced above, the answer seems to be that the subordination complaint is effectively countered once algorithmic rule is introduced. How could people have a subordination complaint against algorithmic systems? By removing people with power and authority, and substituting them with algorithmic systems, political authority ceases to be a problem of relational equality. Surely we are still subject to power and authority, but not to power and authority of other people – which, on the subordination view, is what is wrong with political authority. Substituting those people, it seems, successfully eliminates the problematic relation. I argue that this line of argument however misrepresents what animates the subordination complaint. Substituting human rule with algorithmic rule ideally makes rule better because the less space there is for human discretionary power, the more predictable and less discriminating those rules are. But better and more predictable rule does not eliminate the social hierarchies that animate the subordination complaint: it does not change the fact that people exercise power over other people. Instead, it merely limits the ways in which they can wield power over other people, namely, by making the ways in which power is wielded dependent on algorithmic rules. But it does not remove those who deploy algorithmic systems from that power.