Is it possible that robots like humans belong to a non-human group?

抄録

The uncanny valley phenomenon is assumed to be due to the anxiety of threat to human distinctiveness. According to previous studies, it is hypothesized that human-like robots are considered as belonging to the out-group, which leads to anxiety. Because the human faces of out-group members are difficult to recognize, we predicted that a similar effect would be obtained. The experiment contained a 2 (category: Human vs Human-like Robot) × 3 (facial expression: positive, negative and control) within-subject factor. Sixty face images with different expressions were selected through an online survey. In the experiment, observers learned half of the face images randomly. Then after a 5 minutes crossword puzzle, observers saw all the face images and recognized whether it has been learned before. The results showed that recognition performance was better in the negative expression condition than in the other conditions, although the performances were not significantly different between the categories.