An Exploratory Investigation of the Self Serving Biases of Interface Agent Users

This research presents a preliminary investigation of the self-serving biases of interface agent users. For that, an experiment of current and potential users of Microsoft interface agents was conducted. The analysis demonstrates that interface agent users do not always attribute the successful outcomes of human-agent interaction to themselves and negative results to interface agents. In fact, people’s attributions of credits and responsibilities depend on the category of support provided by an agent. This finding contradicts attribution theory. In addition, the study discovers the positive relationship between the degree of an agent’s autonomy and the presence of self-serving biases. Overall, the research attempts to outline the factors that lead to the self-serving biases of interface agent users and suggests several theoretical and practical contributions for agent designers.