The design of neutrality: Evaluations and biases in the perception of male, female, and neutral avatar faces

This study investigates how participants perceive avatars designed to evoke male, female, or gender-neutral impressions. While previous research has shown that users often project gendered stereotypes onto artificial agents, few studies have directly examined how neutral designs are perceived across diverse gender identities. Our aim is to explore whether gender-neutral avatars elicit more balanced evaluations and reduce the activation of stereotypical associations in human–agent interaction. To do so, we conducted an experiments involving participants identifying as women, men, and non-binary. All participants evaluated avatars created and previously validated to represent feminine, masculine, and neutral appearances. In the experiment, we adopted a mixed design to increase statistical power and generalizability. Participants evaluated a set of three avatars per gender category (female, male, neutral), allowing within-subject comparisons. Additionally, individual differences in gender stereotypes were measured to assess how these beliefs influenced participants’ responses to avatar gender.