On How Agents Make Friends: Mechanisms for Trust Acquisition

Abstract

We need models of trust to facilitate cooperation in multi-agent systems, where agents, human and artificial, do not know each other beforehand. This paper lists and proposes simple mechanisms for trust acquisition based on a very basic and general definition of trust, making no assumptions on the internal cognitive models of the involved agents. We also show how trust acquired one-on-one can be propagated in a social network of agents.