Just AI? Gender, power, and intersectional discrimination

The promise of accuracy, efficiency, and speed, unshackled from human bias and error, informs a widespread narrative that propels the expansion of artificial intelligence (AI) systems across a broad range of domains, from education to health, manufacturing, finance, workplace recruitment and management, even warfare. However, despite AI’s beneficent impact in areas such as disease diagnosis and the optimisation of manufacturing processes, this positive narrative has significant blind-spots with respect to AI’s effects on minoritized groups and populations, including women, ethnic minorities, and LGBTQI+ people. The chapter focuses on the gendered effects of AI, specifically on AI bias and discrimination. Analyzing AI through a socio-technical lens, which conjoins AI’s technical capabilities with the social and political contexts of its design, development, and deployments, the chapter demonstrates how AI systems impact differently—and unequally—on the (working) lives of men and women. Such a socio-technical perspective foregrounds underlying structures of power and inequality, which propel AI harms and injustices, and it interrogates a political rationality that impels the use of AI systems and computational methods in complex social contexts.

The chapter commences with an exposition of the concept of gender and sketches AI’s impact on the study of gender, followed by an analysis of the unequal gendered consequences of the use of AI systems in recruitment practices. It proceeds to discuss the lack of gender diversity in the AI workforce and problematizes the pipeline framework, whilst the final section maps emerging regulatory regimes and argues for the importance of effective and enforceable AI regulation in tackling AI’s gender bias and ensuing practices of gender-based discrimination.