Algorithmic Discrimination and Social Justice from the Perspective of Value Philosophy

Abstract

Algorithms are strategy mechanisms and run programs based on data as the basic resource, aimed at solving problems and completing tasks. They serve as the “nervous centralis” and “soul” of artificial intelligence (AI), but they are not “value-neutral.” When data is collected, stored, and analyzed with technologies such as machine learning for automated decision-making, the standpoint of the one who designs and writes algorithms, the source and accuracy of data supporting and training the algorithms, and the value load and value selection intrinsic to the algorithms, as well as the behavioral tendency of specific value subjects based on the algorithms and the independent evaluation and decision-making of intelligent systems, may all lead to a certain degree of algorithmic discrimination, ultimately affecting the realization of social justice. Compared with the social discrimination familiar to people, algorithmic discrimination is more extensive, diversified, accurate, targeted, concealed, and “cunning.” As we advance the intelligent transformation of society, it is essential to establish the status of all people as value subjects, “embed” the values of justice into intelligent algorithms, and establish dynamic evaluation and supervision mechanisms. Only through these actions can we make necessary regulations on algorithmic discrimination and reconstruct a just social order in the era of intelligence.

Keywords

artificial intelligence (AI) / automated decision-making / algorithmic discrimination / value subject / social justice

RIGHTS & PERMISSIONS

Higher Education Press