The Quagmire Model: How a small step can trap you in the quagmire of misinformation, hate speech and denialism

Abstract

Abstract This study aims to explain how social media (SM) users, whilst searching for information, can be trapped into a quagmire of misinformation, even when they have no denialist inclinations or sympathy for hate groups. We analyze the interactions between cognitive biases and deep preference learning algorithms (DPL), as SM companies use DPL to curate the content conveyed to its users. The study proposes a model for users’ behavior and explain how SM business model allows new information to be introduced in the quagmire in order to change user’s opinions in a way desired by a customer willing to pay for, and, eventually, accomplished it. The model explains why some popular tactics against misinformation, as censorship and fact-checking, achieve very poor results. We suggest that policies promoting face-to-face encounters in friendly environments can be more effective in that struggle. We believe the model can help decision makers in developing more efficient anti-disinformation policies.