| Abstract | We study users' incentives to become cybercriminals when network security is interdependent. We present a game-theoretic model in which each player (i.e., network user) decides his type, honest or malicious. Honest users represent law-abiding network users, while malicious users represent cybercriminals. After deciding on their types, the users make their security choices. We will follow [29], where breach probabilities for large-scale networks are obtained from a standard interdependent security (IDS) setup. In large-scale IDS networks, the breach probability of each player becomes a function of two variables: the player's own security action and network security, which is an aggregate characteristic of the network; network security is computed from the security actions of the individual nodes that comprise the network. This allows us to quantify user security choices in networks with IDS even when users have only very limited, aggregate information about security choices of other users of the network. |
Becoming Cybercriminals: Incentives in Networks with Interdependent Security