Key points are not available for this paper at this time.
Nowadays, most systems use artificial intelligence algorithms to automate tasks and reduce the time required for execution. Moreover, it must estimate the bias risks that can be introduced within the system. Based on these considerations, quantitative measures and prioritization strategies can be established for those inadequate situations, choosing an appropriate method to overcome gender bias. In this study, the impact of gender bias on an annual salary risk score due to gender bias was analyzed to identify and reduce it as much as possible in machine learning algorithms and on text data provided to a virtual assistant. The study finds that gender bias can influence our decisions by illustrating hypotheses on how algorithms affect prioritization decisions and strengthen stereotypes by favoring men against women. Recommendations to lower gender bias can include training programs for poor people that face substantial barriers to accessing education; training programs for people with a low level of education or no access; access to all kinds of jobs for women; assurance of diversity and inclusiveness; and algorithms that are fair and trained with the definite goal of reducing gender bias.
Mercioni et al. (Thu,) studied this question.