Algorithmic reputation and information equity

Abstract

ABSTRACTIn this paper I offer a theory of reputation and information systems. As a form of human judgement and a tool for decision making, I argue that the social process of reputation plays a central role in essential human functions and that one's ability to control reputation is a mark of power over important decisions concerning one's fate. Algorithmic reputation can be understood as a type of reputation that has been transformed from a familiar and visible interpersonal process to a commodified and remote data process. I argue that we ought to evaluate algorithmic reputation as the digital expression of the social process of reputation and situate it within existing moral obligations and interests. Reputations can be fraught with error and bias but perform vital social and economic functions nonetheless. Reputation absolutists would require that we subsume any concerns about reputational fairness, such as the risk of disparate effects on members of marginalized social groups, to the ideals of information transparency and business efficiency. I conclude that we risk enshrining societal inequity as algorithmic destiny by committing to the absolutist position on reputational transparency, posing serious risks to the functioning of a free and fair information society.