AI and Gender Bias in Banking Hiring Practices

Gender bias in hiring is a significant problem in the banking sector. This study explores how individuals’ bias towards hiring males is impacted when those individuals are exposed to existing recommendations with embedded bias towards hiring either males or females. We also investigate whether the impact of these recommendations changes when individuals believe they come from Artificial Intelligence (AI), male partners, or female partners. We perform a 2 × 3 between-subjects experiment that asks subjects to rank candidates under each condition (existing recommendations favor male candidates vs. existing recommendations favor female candidates, and recommendations come from AI vs. recommendations come from male partners vs. recommendations come from female partners). The results show that subjects tend to imitate the male or female favoring bias in the existing recommendations. Results further show that male and female subjects tended to ignore a bias towards hiring males when it came from the opposite sex. Further, subjects with positive perceptions toward AI were more likely to favor male candidates, reflecting the dynamic of perception of credibility, gender bias, and social identity. These results underline the importance of collaboration between AI developers and HR to mitigate biases and address systemic disparities in recruitment in the banking sector.