ABSTRACT
Our field trial highlighted the important unsolved aspects of human-robot interaction in an authentic social setting. The trial showed gradual loss of interest in interacting with the robot among most of the children. It was an important finding that the children interacted with the robot for the duration of 1 week; however, our robots failed to keep most of the children’s interest after the 1st week. We believe that the robots’ first impact created unreasonably high expectations in the children. The children mobbed the robot, overwhelming its ability to interact. In other words, the robot could not cope with the children’s enthusiasm. Although partner robots are making news in Japan (such as Honda’s and Sony’s humanoid robots, and the big exhibition on partner robots named ROBODEX; ROBODEX Executive Committee, 2003), the robots’ ability to be a partner to people is still lacking. Robots are very novel in general; therefore, their first impact can induce a greater desire for communication than their interactive ability can satisfy. In our trial, the children’s interaction with the robots gradually decreased, especially during the 2nd week. Therefore, our trial showed us the limitation of the robots’ ability to maintain long-term relationships and the disappointment that followed the robot’s initial impact. However, we believe unreasonable expectations will diminish as partner robots become commonplace.Regarding the body and appearance of the robot, our results seem to encourage the use of a humanoid robot. We believe that the body of a humanoid robot played a useful part in establishing common ground. That is, a robot that possesses a humanoid body will be more successful at sustaining interaction because people see it as similar to themselves and that it interacts as they do. Nonetheless, we need further research to establish a model of these kinds of social effects, such as common ground to see if they are more easily achieved with a humanoid robot by comparing humanoid and nonhumanoid robots.