On detecting deception in agent societies

Driven by the advances in intelligent agent technologies, agent societies have taken on prominent roles in various sciences and applications. Research on risks posed to such societies such as deception by malicious agents has recently begun gaining ground. Drawing on the deception literature from social sciences and artificial intelligence, we propose a deception model which extends and integrates the past research on deception from multiple disciplines. In order to facilitate deception detection, we also develop a procedural model for deception detection. Application of both models to a real-world case scenario demonstrates that both models can be used to instill a better understanding of deception and to detect and remedy deception effectively.