Synergies Between Deplatforming and Inoculation Against Online Extremism: An Agent-Based Simulation and Cost-Benefit Analysis

• In Event: The Mechanics of Online Hate: Understanding and Disrupting Digital Pathways to Extremism

Abstract

Inoculation, (i.e., exposing users to weakened forms of extremist arguments to build belief resistance) is known to effectively reduce the spread of online extremism. However, it is unclear to what extent inoculation is needed when combined with other counter-extremism strategies (e.g., deplatforming of extremist users). We develop an agent-based model of diffusion and control of online extremism and calibrate it using data from systematic reviews and empirical studies. A population of agents forms a binary belief about polarizing content unilaterally spread by extremists embedded within the social network. We introduce two interventions to mitigate the spread of extremism: deplatforming and inoculation. We find strong evidence that inoculation is synergetic to deplatforming and pin down this relationship to a protective effect exerted by inoculation on individual belief formation, which, at the group level, translates into a stabilization of opinion dynamics. Hence, the introduction of inoculation on top of deplatforming makes the outcome of a deplatforming intervention less arbitrary on average.
We exploit the partial substitution effect to identify the minimum amount of inoculation required to control the spread of extremism for each additional dollar spent on repression. Cost-benefit analysis highlights the efficiency advantages of having a comprehensive policy approach encompassing both measures.

Authors

• Matteo Vergani, Deakin University

• Andrea Giovannetti, Australian Catholic University

• Stephanie Zi Xin Ng, Deakin University

• Chee Peng Lim, Deakin University

• James Zhang, Deakin University

• Robin Scott, Deakin University