Gender Encoding in Industrial Robots: All Work, No Play, and No Significant Differences

Encoding gender in robots is a common practice, especially in social contexts where robots usually have a higher degree of anthropomorphic design. However, even industrial robots, traditionally having a lower degree of anthropomorphic design, are sometimes imbued with gender cues, like male or female names or stereotypical colors. This practice persists despite limited research exploring the implications of gender encoding in industrial settings. Our study explored the effects of encoding gender cues in an industrial robot on stereotypical perceptions, trust, and behavioral dependence. In a laboratory experiment, participants performed a quality assurance task for welding seams with a collaborative robot encoded as female, male, or non-gendered through name, pronouns, and color. The robot provided scans of potentially faulty weld seams along with an analysis of the proportion of defective areas. Participants then offered their estimates of the faulty areas. We measured gender-stereotypical perceptions, including warmth, communion, agency, and competence, as well as subjective trust and behavioral dependence, assessed by the extent to which participants deviated from the robot's analysis. Interestingly, the most notable finding was the absence of significant effects of encoded robot gender. Contrary to expectations, no significant differences in perceived warmth, agency, competence, subjective trust, or behavioral dependence were found across conditions. However, the male-encoded robot was rated as more communal and transparent, measured as part of a multidimensional trustworthiness assessment, than the non-gendered robot, though these findings warrant replication. Overall, the results support a growing body of research suggesting that encoding robot gender does not provide interactional benefits. Importantly, our findings do not imply that gender cues should be used if found advantageous - in fact, intentionally designing robots with gendered features risks reinforcing harmful stereotypes. Instead, our study highlights that minimizing gender cues can sustain effective interactions in industrial HRI without perceptual or behavioral disadvantages.