IN THE PRESENCE OF HUMAN-LIKENESS An Exploratory Study on How Humans Experience Artificial Agents

Abstract: Interactions with artificial agents are increasing in everyday situations, and social agents might become the next generally adopted interface. However, whether these agents should appear human-like or not is debated as it is still uncertain what effects human-like agents have on humans in an interactive setting. There are several aspects that can affect human responses to artificial agents, such as physical presence and expectations. The uncanny valley hypothesis has been influential in explaining human reactions to human-like agents, but the validity of this hypothesis is still questioned. Earlier studies have mainly been conducted using quantitative measures, which do not seem to capture the whole human experience, hence a wider perspective is needed. Virtual and physical agents might become available in different situations due to financial or technical reasons. Because they differ in physical presence they may be perceived and experienced differently. With a qualitative approach this study explored two scenarios, a physical and a virtual one, using the social robot Furhat. The participants met one version of Furhat in an interactive session. Data was collected from a pre-interaction survey, facial emotion recognition and interviews that was thematically analysed. Due to the study’s exploratory approach no generalizations of the results could be made, instead topics for future research were identified. Suggestions from this study is to further investigate feelings of uncanniness in human-robot-interaction, habituation processes, relationships between physical and social presence, and when it is suitable to use human-like design.