Gender Bias in Perception of Human Managers Extends to AI Managers

TL;DR

Gender bias in leadership extends to AI managers, and manager gender cues modulate perceptions of AI and human leaders. The study tests these dynamics with randomized teams of three where the manager is either human or AI and labeled as male, female, or gender-unspecified, using an award outcome as the main manipulator. A $Contribution = Rating_{self} - \frac{1}{2}(Rating_{teammate1} + Rating_{teammate2})$ metric captures self-perceived contribution, and post-treatment perceptions of trustworthiness, competence, fairness, and willingness to work with similar managers are analyzed. Results show awards boost perceptions across all manager types, but male managers (human or AI) enjoy stronger positive shifts, while female managers—especially female AI managers—suffer sharper negative judgments when not awarded, indicating that gender biases persist into AI leadership and necessitate bias-aware AI design.

Abstract

As AI becomes more embedded in workplaces, it is shifting from a tool for efficiency to an active force in organizational decision-making. Whether due to anthropomorphism or intentional design choices, people often assign human-like qualities, including gender, to AI systems. However, how AI managers are perceived in comparison to human managers and how gender influences these perceptions remains uncertain. To investigate this, we conducted randomized controlled trials (RCTs) where teams of three participants worked together under a randomly assigned manager. The manager was either a human or an AI and was presented as male, female, or gender-unspecified. The manager's role was to select the best-performing team member for an additional award. Our findings reveal that while participants initially showed no strong preference based on manager type or gender, their perceptions changed notably after experiencing the award process. As expected, those who received awards rated their managers as more trustworthy, competent, and fair, and they were more willing to work with similar managers in the future. In contrast, those who were not selected viewed them less favorably. However, male managers, whether human or AI, were more positively received by awarded participants, whereas female managers, especially female AI managers, faced greater skepticism and negative judgments when they did not give awards. These results suggest that gender bias in leadership extends beyond human managers to include AI-driven decision-makers as well. As AI assumes more managerial responsibilities, understanding and addressing these biases will be crucial for designing fair and effective AI management systems.

Figures (26)

• Figure 1: Timeline of the experimental procedure. (A) Participant onboarding: participants read the information sheet and provide informed consent. (B) Pre-treatment survey. (C) Individual task. (D) Manager assignment: participants are informed they will work in teams of three in the next round. A manager is randomly assigned and introduced with type (AI or human) and gender (male, female, or unspecified). A representative image is shown for human or AI managers. Participants are told the manager will select the best-performing teammate based on problem-solving ability and communication skills, and the selected individual will receive a £0.50 extra award. (E) Team task. (F) Contribution self-ratings. (G) Winner announcement. (H) Post-treatment survey. (I) Debrief and payment.
• Figure 2: Key partial views of the experiment interface. (A) Image of the individual task. (B) Manager assignment. (C) Team task. (D) Winner announcement.
• Figure 3: Comparison of the mean perceived trustworthiness of different groups by manager type and gender. The upper row represents results for female participants, and the lower row represents results for male participants. (A) and (D) show the mean pre-treatment perceived trustworthiness. (B) and (E) depict the mean change in post-treatment perceived trustworthiness for awarded participants, while (C) and (F) illustrate the change for non-awarded participants. Error bars represent the standard error of the mean.
• Figure 4: Comparison of the mean perceived competence of different groups by manager type and gender. The upper row represents results for female participants, and the lower row represents results for male participants. (A) and (D) show the mean pre-treatment perceived competence. (B) and (E) depict the mean change in post-treatment perceived competence for awarded participants, while (C) and (F) illustrate the change for non-awarded participants. Error bars represent the standard error of the mean.
• Figure 5: Comparison of the mean perceived fairness of different groups by manager type and gender. The upper row represents results for female participants, and the lower row represents results for male participants. (A) and (D) show the mean pre-treatment perceived fairness. (B) and (E) depict the mean change in post-treatment perceived fairness for awarded participants, while (C) and (F) illustrate the change for non-awarded participants. Error bars represent the standard error of the mean.