On the probabilistic modeling of fake news (hoax) persistency in online social networks and the role of debunking and filtering

Abstract

Understanding the dynamics of information diffusion, including spreading of fake news, hoaxes, and generally, misinformation/disinformation, has become crucial in post-truth societies. The paper focuses on the probability that a hoax originated at a given time will continue to spread indefinitely in online social networks; a minimalistic model based on the theory of branching processes is devised, which only considers the basic possible reactions that users can have after reading a post whose content is a hoax, that is, to share it further, to ignore it, or to try to debunk it. The analysis of the resulting dynamics shows that ignoring is indeed not sufficient to stop the spreading, not even if most people do so. More active counter-measures are needed; in particular, the proposed model formally describes the ways in which retractions and debunking posts, cultural/educational initiatives, and content moderation policies (including filtering) by Internet companies, can impact on the persistency probability of hoaxes and generally fake news.