Gender Bias in AI: Implications for Managerial Practices

Publication Type:
Conference Proceeding
Responsible AI and Analytics for an Ethical and Inclusive Digitized Society, 2021, 12896, pp. 259-270
Issue Date:
Filename Description Size
Nadeem2021_Chapter_GenderBiasInAIImplicationsForM.pdfPublished version231.07 kB
Adobe PDF
Full metadata record
Artificial intelligence (AI) applications are widely employed nowadays in almost every industry impacting individuals and society. As many important decisions are now being automated by various AI applications, fairness is fast becoming a vital concern in AI. Moreover, the organizational applications of AI-enabled decision systems have exacerbated this problem by amplifying the pre-existing societal bias and creating new types of biases. Interestingly, the related literature and industry press suggest that AI systems are often biased towards gender. Specifically, AI hiring tools are often biased towards women. Therefore, it is an increasing concern to reconsider the organizational managerial practices for AI-enabled decision systems to bring fairness in decision making. Additionally, organizations should develop fair, ethical internal structures and corporate strategies and governance to manage the gender imbalance in AI recruitment process. Thus, by systematically reviewing and synthesizing the literature, this paper presents a comprehensive overview of the managerial practices taken in relation to gender bias in AI. Our findings indicate that managerial practices include: better fairness governance practices, continuous training on fairness and ethics for all stakeholders, collaborative organizational learning on fairness & demographic characteristics, interdisciplinary approach & understanding of AI ethical principles, Workplace diversity in managerial roles, designing strategies for incorporating algorithmic transparency and accountability & ensuring human in the loop. In this paper, we aim to contribute to the emerging IS literature on AI by presenting a consolidated picture and understanding of this phenomenon. Based on our findings, we indicate direction for future research in IS for the better development and use of AI systems.
Please use this identifier to cite or link to this item: