A Generic Model of Distrust Behaviour in Online Communities of Trust

With boost of interest in multi-user networking ranging from Web 2.0 technologies to agent systems and smart environments, various aspects of mutual interactions between system components, including but not limited to human users themselves, have to be considered. Trust is one of core dimensions of vital, multilaterally beneficial interactions. In the first section of the paper we describe the trust itself and its characteristics, such as multidimensionality, contextuality, asymmetry, transitivity, scope, disproportion, and dynamics. Further we present a niche type of community where users trust each other as default and where the trust loses most of its subjective flavour, which we call a community of trust. Common trust models do not apply for the community of trust well, because they build upon different presumptions about trust. As a main contribution of the paper, we introduce a simple generic and extensible model of distrust for the community of trust. The model is based on activity and distrust matrices arranged into vectors in order of time of occurrence. Various derived characteristics, such as harmful/harmless or distrusting/distrusted ratios provide additional insight on the model. Tests in simulated scenarios and with real human users will follow in our future research in order to concretize and evaluate the model.