Most of current social network services are vulnerable to malicious actions. For example, rumor (e.g., contaminated and distorted information) can be diffused along the social links. In this paper, given a social network service, we design a peer-to-peer (P2P) network, and propose a robust information diffusion model to efficiently detect the malicious peers from which a risk (i.e., rumor) has been generated on the P2P network. Thereby, by aggregating social interactions among users, a set of interaction sequences are obtained. Given a set of interaction sequences, statistical sequence mining method is exploited to discover a certain social position which provides peculiar patterns on the P2P networks. For evaluating the proposed method, we conducted two experimentations with NetLogo simulation platform for risk discovery on social network