Presentation
From Screen to Scene: Using Video vs. Virtual Reality in Human-Robot Visual Perspective Taking
SessionLBR3: Late-Breaking Results
DescriptionVisual perspective-taking (VPT), considering another's viewpoint, is crucial for social interaction. While research suggests humans engage in VPT not only in human-human but also during human-robot interaction, studies have primarily used static images or videos of robots to investigate this phenomenon. To investigate if the results hold for co-present interaction partners, we conducted a study investigating whether VPT depends on the medium the robotic avatar is presented via (Video vs. VR). 84 participants (aged 18-29) were randomly assigned to the video or VR condition. Participants had to indicate what number they had seen on the table between them and the robot (6 or 9). Consequently, we calculated the proportion of participants who adopted the robot’s viewpoint as the VPT rate in each condition. Moreover, we also measured feelings of being with the robot (i.e., perceived co-presence). Surprisingly, the data analysis showed that VPT rates in the video condition were significantly higher than in the VR condition, even though the perceived co-presence was, as expected, higher in the VR condition. These unexpected findings might have uncovered a general misconception in VPT in earlier studies where the stimuli (6 vs. 9) were co-present in the robot's world (picture or video) and not the human’s environment, which could have led to the spontaneous VPT. The results suggest that a potential lack of willingness to adopt the perspective of robots underscores the increased importance of exploring under what circumstances humans are willing to adopt the visual and possibly affective perspective of robots.
Event Type
Late Breaking Results
TimeThursday, September 12th3:20pm - 3:30pm MST
LocationFLW Salon D