Algorithmic fairness and deliberative self-determination

Abstract

If democracy is about enabling all people to have equal opportunities to influence the decisions that affect them, digital societies need to ask how to ensure that new environments make this equality feasible. The first challenges are conceptual: understanding how the interaction between humans and algorithms is configured, what the learning of these devices consists of, and the nature of their biases. Immediately afterwards, we come up against the unavoidable question of what kind of equality, we are trying to ensure, bearing in mind the diversity of conceptions of fairness in our societies. If articulating this pluralism is not a matter that can be resolved with an aggregative technique, but requires political compromises, then a deliberative conception of democracy seems the most apt to achieve the equality to which democratic societies aspire.