Examining Trust, Forgiveness and Regret as Computational Concepts

Abstract

The study of trust has advanced tremendously in recent years, to the extent that the goal of a more unified formalisation of the concept is becoming feasible. To that end, we have begun to examine the closely related concepts of regret and forgiveness and their relationship to trust and its siblings. The resultant formalisation allows computational tractability in, for instance artificial agents. Moreover, regret and forgiveness, when allied to trust, are very powerful tools in the Ambient Intelligence (AmI) security area, especially where Human Computer Interaction and concrete human understanding are key. This paper introduces the concepts of regret and forgiveness, exploring them from social psychological as well as a computational viewpoint, and presents an extension to Marsh’s original trust formalisation that takes them into account. It discusses and explores work in the AmI environment, and further potential applications.