Agent-based simulation of fake news dissemination: the role of trust assessment and big five personality traits on news spreading

The recent development of Social Networking Services (SNS) has changed the way of information sharing and interchanging. With countless amounts of information broadcast daily, many people use it every time, even during a crisis and disaster situation. However, the trustworthiness of information has become one of the issues on SNS. In this paper, we propose a trust model consisting of identity-based, behavior-based, relation-based, feedback factors, and information-based trust, in which the Big-Five personality traits are also considered. We conducted an agent-based modeling simulation for the proposed trust model, investigating users’ behavior according to the Big-Five personality traits and several users’ aspects: knowledge level and psychopathy. The experiment is based on online surveys and related works representing social network users’ behavior. We compare the overall trust and trustworthiness in the numerical results, validating our proposed trust model. Moreover, we systematically compare the occurrence of fake news under conditions where the initial news is either a truthful or fabricated fake news. Numerical results also show that overall trust is sensitive to information-based trust, whereas it is not affected by behavior-based trust. Furthermore, openness, conscientiousness, and extroversion were correlated with overall trust, while the effect of agreeableness and neuroticism on overall trust were insignificant.