Algorithmic Biases and Social Media

Abstract

This final chapter of Part II of the book is a bit different from the previous chapters insofar as it zooms into one particular situation where algorithmic bias occurs: the choice of posts shown to social media users. In doing so, I achieve two objectives: this serves as a case study that shows how the various biases discussed so far can interact and reinforce each other, and it illustrates how algorithmic bias can be dynamic. Rather than set in stone, the bias can develop and grow over time out of an interaction between user and algorithm—the biases of both reinforce each other until the result can be a discomfortingly strong distortion.