Affective Interfaces of Embodied Conversational Agents

- A Study of Character Interfaces -

Abstract

The visual representations of intelligent agents draw increasing attention recently. Additionally, the first impression of humans toward agents is the visual representations. How these man-made creatures' looks determine users' will to interact with. This research focuses on the appearances of embodied agents to investigate the affective interactions between users and agents. This study conducted one experiment to study the affective influences of embodied conversational agents (ECAs) on users in the learning tasks. A one-factor-at-three-levels counterbalanced by within-subjects design is employed in this study. This study asks whether there are significant experiential differences between ECAs when they are represented by different character classifications, such as character preference, user engagement and user-agent relationships. The main contributions of this study are summarised as follows: (1) character classifications can be related to human affective factors; and (2) there may be significant relations among character preference, engagement and user-agent relationships.