As social networking sites continue to proliferate, online deception is becoming a significant problem. Deceptive users are now not only lone wolves propagating hate messages and inappropriate content, but also are more frequently seemingly honest users choosing to deceive for selfish reasons. Their behavior negatively influences otherwise honest online community members, creating a snowball effect that damages entire online communities. In this paper, we study the phenomenon of deception and attempt to understand the dynamics of users’ deception, using a game-theoretic approach. We begin by formulating the decision process of a single user as a Markov chain with time-varying rewards. We then study the specific optimization problem a user may face in choosing to deceive when they are influenced by (1) their potential reward, (2) peer pressure and (3) their deception comfort level. We illustrate reasonable equilibria can be achieved under certain simplifying assumptions. We then investigate the inverse problem: given equilibria, we show how we can fit a model to the data and how this model exposes information about the social structure.
All Science Journal Classification (ASJC) codes
- Information Systems
- Media Technology
- Human-Computer Interaction
- Computer Science Applications