Challenges for Trust, Fraud and Deception Research in Multi-agent Systems

Discussions at the 5th Workshop on Deception, Fraud and Trust in Agent Societies held at the 1st International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS 2002) centered around many important research issues1. This paper attempts to challenge researchers in the community toward future work concerning three issues inspired by the workshop’s roundtable discussion: (1) distinguishing elements of an agent’s behavior that influence its trustworthiness, (2) building reputation-based trust models without relying on interaction, and (3) benchmarking trust modeling algorithms. Arguments justifying the validity of each problem are presented, and benefits from their solutions are enumerated.