UNFAIR ALGORITHMS. MAPPING THE CURRENT RESEARCH LANDSCAPE ON GENDER BIASES IN ALGORITHMIC SYSTEMS

Algorithms wield significant influence in culture and society, impacting choices, behaviours, and collective imaginaries (Manovich, 2018; Martin and Mason, 2023). The role of technology platforms in shaping content exposure and online flourishing has sparked considerable debate across different fields (WEF, 2018), with a recognition that algorithmic systems can heighten gender stereotypes and biases (Gupta et al., 2021) and diminish aesthetic diversity over time (Manovich, 2018; Rementeria-Sanz, 2020). This paper provides an overview of findings emerging from a preliminary analysis of academic publications about the relationship between algorithms and gender and ethnic biases, that have been carried out within a wider research project aimed at investigating algorithmic systems through the lens of gender and ethnicity. Two perspectives have been considered: sociology, examining aspects like human-computer interaction, word embedding processes, image processing, classification, and user perceptions; and computer science, focusing on strategies to mitigate gender biases in algorithmic systems. Examining the growing literature around the area of gender and ethnicity biases in algorithms, it is possible to highlight that there is a pressing need to scrutinise how algorithmic systems contribute to existing patterns of inequality (Eubanks, 2018). Indeed, our analysis underscores the necessity for future research to delve into issues surrounding the fairness of algorithmic systems, including the embedded conceptions of gender in technological design, the use of biased-datasets amplifying gender and ethnicity inequality, the role of professionals in the field, and users’ perceptions - an equally crucial yet understudied aspect (Wang et al., 2022).