The Case of K.I.T.T. and Data – from Science Fiction to Reality? A Social Psychology Perspective on Artificial Companions

With the development of companion systems, research on virtual agents and robots gains increasing attention in the media and is brought to the public´s focus. Indeed, social science research and public discussion on current developments is necessary since the implications can be discussed controversially. Do people want to share their bed and board with an artificial significant other? In Science Fiction, companion technologies are part of the daily lives of the protagonists. Michael Knight, for instance, has been teamed up with the robotic car K.I.T.T. and Commander Data is a well-respected member of the USS Enterprise crew. While these Sci-Fi companions are perfectly designed systems users are happy to deal with, in reality researchers and developers face the frequently occurring phenomenon that people are initially interested in interacting with an artificial entity; but are, however, quickly bored or annoyed with it, refuse to use it again and even show aggression towards the system (de Angeli et al. 2006; Walker et al. 2002). Regarding service and assistive technologies, long term interaction is yet particularly important, because these systems can unfold their full potential only over a certain period of time. Thus, the central challenge is to further refine the sociability of artefacts that is considered to facilitate human-robot/agent interaction (HRI/HAI; Kramer et al. 2011). The second major challenge is to choose and use adequate methods to study human-companion relationships. So far, only few long-term studies with virtual and robotic companions have been conducted and most of them are restricted to very simple methodologies.  This paper will thus give an overview on existing research on companions in HCI and HRI, discuss the applicability of the underlying theoretical assumptions on the sociability of artefacts and provide an overview and discussion of methods used for artificial companion research.