Search results

Gender bias in artificial intelligence (AI) should be solved as a priority before AI algorithms become ubiquitous, perpetuating and accentuating the bias. While the problem has been identified as an established research and policy agenda, a cohesive review of existing research specifically addressing gender bias from a socio-technical viewpoint is lacking. Thus, the purpose of this study is to determine the social causes and consequences of, and proposed solutions to, gender bias in AI algorithms.

A comprehensive systematic review followed established protocols to ensure accurate and verifiable identification of suitable articles. The process revealed 177 articles in the socio-technical framework, with 64 articles selected for in-depth analysis.

Most previous research has focused on technical rather than social causes, consequences and solutions to AI bias. From a social perspective, gender bias in AI algorithms can be attributed equally to algorithmic design and training datasets. Social consequences are wide-ranging, with amplification of existing bias the most common at 28%. Social solutions were concentrated on algorithmic design, specifically improving diversity in AI development teams (30%), increasing awareness (23%), human-in-the-loop (23%) and integrating ethics into the design process (21%).

This systematic review is the first of its kind to focus on gender bias in AI algorithms from a social perspective within a socio-technical framework. Identification of key causes and consequences of bias and the breakdown of potential solutions provides direction for future research and policy within the growing field of AI ethics.

The peer review history for this article is available at https://publons.com/publon/10.1108/OIR-08-2021-0452

In response to calls to reduce the gender gap in the salesforce, this study aims to examine the effect of candidate gender, manager gender and industry to explain gender bias in salesperson recruitment during screening and skill assessment.

This paper tested the hypotheses using observational data from a national sales competition in the USA, where managers evaluated student candidates for entry-level sales positions.

This research finds gender bias during screening using the dyadic perspective. Specifically, female managers evaluate male candidates more favorably than male managers do during screening. Further, managers of service companies evaluate female candidates more favorably than managers of goods companies during screening. However, this paper finds no such effects during candidates’ skill assessment.

The findings indicate the importance of using dyadic research techniques to assess gender bias.

Managers should not use short interactions to screen candidates.

Implicit bias exists when candidates and managers interact during screening. To reduce gender bias in recruitment the candidates and managers should interact for a longer duration.

This study draws upon a unique setting, where the candidates interact with the managers for screening and skill assessment. Implicit bias exists when candidates and managers interact for screening under time pressure. This paper finds no evidence of gender bias in skill assessment. This study finds that female managers are more prone to bias when evaluating male candidates than male managers. Prior work has not examined industry-based bias; this paper provides evidence of such bias in candidate screening.

Drawing on attribution theory, this study aims to examine how and when abusive supervision affects insubordination, focusing on employees’ attribution bias related to leader gender.

Two mixed-method studies were used to test the proposed research framework. Study 1 adopted a 2 (abusive supervision: low vs high) by 2 (leader gender: male vs female) by employee gender-leadership bias quasi-experiment. A sample of 173 US F&B employees completed Study 1. In Study 2, 116 hospitality employees responded to two-wave, time-lagged surveys. They answered questions on abusive supervision and gender-leadership bias in Survey 1. Two weeks later, they reported negative external attribution (embodied in injury initiation) and insubordination.

Hayes’ PROCESS macro results verified a three-way moderated mediation. The three-way interaction among abusive supervision, leader gender and gender-leadership bias affects external attribution, increasing insubordination. Employees with high leader–gender bias working under female leaders make more external attribution and engage in subsequent insubordination in the presence of abusive supervision.

This study is one of the first, to the best of the authors’ knowledge, that examines the mediating role of external attribution of abusive supervision. Second, this research explains the gender glass ceiling by examining employees’ attribution bias against female leaders.

This paper compares how women leaders in four US industries–higher education, faith-based non-profits, healthcare and law–experience 15 aspects of gender bias.

This study used convergent mixed methods to collect data from 1,606 participants. It included quantitative assessment of a validated gender bias scale and qualitative content analysis of open-ended responses.

Results suggest that, while gender bias is prevalent in all four industries, differences exist. Participants in higher education experienced fewer aspects of gender bias than the other three industries related to male culture, exclusion, self-limited aspirations, lack of sponsorship and lack of acknowledgement. The faith-based sample reported the highest level of two-person career structure but the lowest levels of queen bee syndrome, workplace harassment and salary inequality. Healthcare tended towards the middle, reporting higher scores than one industry and lower than another while participants working in law experienced more gender bias than the other three industries pertaining to exclusion and workplace harassment. Healthcare and law were the two industries with the most similar experiences of bias.

This research contributes to human resource management (HRM) literature by advancing understanding of how 15 different gender bias variables manifest differently for women leaders in various industry contexts and by providing HRM leaders with practical steps to create equitable organizational cultures.

The purpose of this paper is to propose that in order to tackle the question of bias in algorithms, a systemic, sociotechnical and holistic perspective is needed. With reference to the term “algorithmic culture,” the interconnectedness and mutual shaping of society and technology are postulated. A sociotechnical approach requires translational work between and across disciplines. This conceptual paper undertakes such translational work. It exemplifies how gender and diversity studies, by bringing in expertise on addressing bias and structural inequalities, provide a crucial source for analyzing and mitigating bias in algorithmic systems.

After introducing the sociotechnical context, an overview is provided regarding the contemporary discourse around bias in algorithms, debates around algorithmic culture, knowledge production and bias identification as well as common solutions. The key concepts of gender studies (situated knowledges and strong objectivity) and concrete examples of gender bias then serve as a backdrop for revisiting contemporary debates.

The key concepts reframe the discourse on bias and concepts such as algorithmic fairness and transparency by contextualizing and situating them. The paper includes specific suggestions for researchers and practitioners on how to account for social inequalities in the design of algorithmic systems.

A systemic, gender-informed approach for addressing the issue is provided, and a concrete, applicable methodology toward a situated understanding of algorithmic bias is laid out, providing an important contribution for an urgent multidisciplinary dialogue.

Women in medicine face barriers that hinder progress toward top leadership roles, and the industry remains plagued by the grand challenge of gender inequality. The purpose of this study was to explore how subtle and overt gender biases affect women physicians, physician leaders, researchers, and faculty working in academic health sciences environments and to further examine the association of these biases with workplace satisfaction. The study used a convergent mixed methods approach. Sampling from a list of medical schools in the United States, in conjunction with a list of each state's medical society, the authors analyzed the quantitative survey responses of 293 women in medicine. The authors conducted ordinary least squares multiple regression to assess the relationship of gender barriers on workplace satisfaction. Additionally, 132 of the 293 participants provided written open-ended responses that were explored using a qualitative content analysis methodology. The survey results showed that male culture, lack of sponsorship, lack of mentoring, and queen bee syndrome were associated with lower workplace satisfaction. The qualitative results provided illustrations of how participants experienced these biases. These results emphasize the obstacles that women face and highlight the detrimental nature of gender bias in medicine. The authors conclude by presenting concrete recommendations for managers endeavoring to improve the culture of gender equity and inclusivity.

The eradication of gender discrimination at work has been a prominent feature of the UK political and business agenda for decades; however, the persistent business gender leadership gap remains. The concept of second-generation gender bias has recently been proposed as the primary cause. This paper aims to evaluate how women experience second-generation gender bias in construction organisations. It examines key manifestations of second-generation gender bias and how it impacts women’s career progression into leadership positions in the UK construction industry.

This paper adopts a broad feminist interpretative lens aligned with the general aims of feminist critical inquiry through semi-structured interviews with 12 women experiencing career journeys of at least five years in the construction industry.

This paper reveals that second-generation gender bias hinders the career development and leadership identity of some women and the persistent business gender leadership gap is unlikely to change without addressing it.

There is little or no research that speaks exclusively to the experience of second-generation gender bias and female managers working within the UK construction. This paper provides further insight into the barriers women face when attempting to progress into senior management roles, particularly in construction.

The purpose of this paper is to examine the effects of gender and strength of gender identification on employees’ perception of gender discrimination. It also explores whether gender comparison and perceived gender bias against women act as mediators in the above relationships. It aims to advance the understanding of the processes leading to individual’s perception of gender discrimination in the Chinese workplace.

Data were collected from 362 workers via an employee survey in three large companies in China. The human resource staff helped us to distribute a self-administered questionnaire to the employees, and the authors assured them of confidentiality and protected their anonymity. To test the hypotheses, the authors employed structural equation modeling. The authors first conducted confirmatory factor analysis on the measurement model, and then the authors estimated three nested structural models to test the mediating hypotheses.

The results reveal that gender and strength of gender identification are related to perceived gender discrimination. The authors further found that gender comparison and perceived gender bias against women partially mediated the relationship between gender and perceived gender discrimination, while gender comparison fully mediated the relationship between strength of gender identification and perceived gender discrimination.

The study helps managers understand why and how their subordinates form perceptions of gender discrimination. Given the findings, they should be aware of the importance of gender identity, gender comparison, and gender bias in organizational practices in affecting such perceptions.

This study is the first exploration of the complex relationships among gender, gender identification, gender comparison, perceived gender bias against women, and perceived gender discrimination. It shows the salient role of gender comparison and gender bias against women in shaping employees’ perceptions of gender discrimination, apart from the direct effects of gender and strength of gender identification.

This paper aims to revisit the extant evidence on gender bias in the market. Specifically, it revisits reaction to CEO announcements. Also, it explores whether the development of the bias over time and by firm size aligns with existing theory.

The paper examines cumulative abnormal returns around CEO announcements from 1992 through 2016 using a modified event study methodology. This evidence shown examines market reactions over time and by firm size.

Financial markets react more favorably to male CEO announcements, with a cumulative abnormal return of 49 basis points above the reaction to their female counterparts. Moreover, the paper finds that market reaction varies over time, which may be because of the increasing proportion of female CEOs, and by firm size, which may be due to the differences in new information available to investors.

Limitations include sample size due to the paucity of female CEO announcements. This paper does not examine the effect of industry, detailed CEO characteristics or announcement content on market reaction. In addition, using an extended event window may increase the likelihood of capturing confounding events, such as mergers or earnings announcements, which limits the interpretability of the results.

Gender bias in financial markets creates another institutional barrier for the advancement of female professionals, as well as implies inefficient capital allocation in markets.

The literature in this field is still inconclusive. Furthermore, bias development over time and the effect of information on bias remain unexplored. This study aims to fill that gap; furthermore, it introduces an extended event-window approach.

The purpose of this paper is to investigate whether machine learning induces gender biases in the sense of results that are more accurate for male authors or for female authors. It also investigates whether training separate male and female variants could improve the accuracy of machine learning for sentiment analysis.

This paper uses ratings-balanced sets of reviews of restaurants and hotels (3 sets) to train algorithms with and without gender selection.

Accuracy is higher on female-authored reviews than on male-authored reviews for all data sets, so applications of sentiment analysis using mixed gender data sets will over represent the opinions of women. Training on same gender data improves performance less than having additional data from both genders.

End users of sentiment analysis should be aware that its small gender biases can affect the conclusions drawn from it and apply correction factors when necessary. Users of systems that incorporate sentiment analysis should be aware that performance will vary by author gender. Developers do not need to create gender-specific algorithms unless they have more training data than their system can cope with.

This is the first demonstration of gender bias in machine learning sentiment analysis.