Reflections on Algorithmic Reputation: Judgment and Equity in a Digitally Mediated Society

Abstract

Contemporary lives are digitally-mediated lives. A significant amount of communication, knowledge-seeking, employment, politics, and transactions take place through devices and on platforms where they are captured, aggregated, and analyzed computationally. This work, often labeled “algorithmic profiling,” is directed toward predicting what people will do and for calculating their potential worth and risk in many domains of life, from policing to employment to dating. Descriptive accounts of algorithmic profiling typically fail to signal how computational judgment and categorization are socio-technical, subject to human culture and politics. I employ reputation as a means of evaluating algorithmic profiling, redirecting attention from technical issues and constraints to its social, political, and economic features. I employ and modify the political philosophy of John Rawls to gain perspective on the moral content of algorithmic reputation. I analyze algorithmic reputation for its role in structuring governing institutions and constructing relations of power. I argue that the requirements of a just, equitable, and stable society include algorithmic reputation processes and practices that are transparent, accountable, and which demonstrate fundamental respect for persons.