(Hu)man-Like Robots: The Impact of Anthropomorphism and Language on Perceived Robot Gender.

Abstract

Implementing anthropomorphic features to robots is a frequently used approach to create positive perceptions in human-robot interaction. However, anthropomorphism does not always lead to positive consequences and might trigger a more gendered perception of robots. More precisely, anthropomorphic features of robots seem to evoke a male-robot bias. Yet, it is unclear if this bias is induced via a male appearance of higher anthropomorphic robots, a general male-technology bias, or even due to language aspects. As the word robot is differently grammatically gendered in different languages, this might be associated with the representation of robot gender. To target these open questions, we investigated how the degree of anthropomorphism and the way the word robot is gendered in different languages, as well as within one language influence the perceived gender of the robot. We therefore conducted two online-studies in which participants were presented with pictures of differently anthropomorphic robots. The first study investigated two different samples from which one was conducted in German, as grammatically-gendered language, and one in English as natural gender language. We did not find significant differences between both languages. Robots with a higher degree of anthropomorphism were perceived as significantly more male than neutral or female. The second study investigated the effect of grammatically-gendered descriptions (feminine, masculine, neuter) on the perception of robots. This study revealed that masculine grammatical gender tends to reinforce a male ascription of gender-neutral robots. The results suggest that the male-robot bias found in previous studies seems to be associated with appearance of most anthropomorphic robots, and the grammatical gender the robot is referenced by.