Filter bubbles as a vector of tradition? Decoding opinion dynamics with agent-based modelling

This paper investigates the relationship between online filter bubbles and the emergence of echo chambers. In particular, building on an agent-based simulation model, we find that the narrower the filter bubble, the larger the number of echo chambers, and the lower the degree of consensus among the population of agents as a whole. According to the existing literature, echo chambers present risks of polarization, extremism and separatism. Yet, we do not recommend reinforcing the regulation of filter bubbles. Indeed, our model suggests that the presence of sources of serendipitous knowledge, like for instance in the form of third places, can prevent the apparition of echo chambers, even in the presence of narrow filter bubbles. Going back to Gabriel Tarde’s classical distinction between opinion, tradition, and reason, we argue that sources of serendipitous knowledge favor the development of a society of opinion, while echo chambers resemble micro-societies of tradition. In this perspective, current developments of information technologies produce a tension between tradition and opinion.