Data Workers in AI development

Abstract

The theme of sovereignty in relation to digital technologies has gained growing attention in recent years, as demonstrated by concepts such as data sovereignty (Hummel et al., 2021) and digital sovereignty (Pinto, 2018). The literature on digital sovereignty comprises a diverse theoretical-methodological universe (Couture, Toupin, 2019), with authors addressing different dimensions of the theme, from the sovereignty of the data generated by groups, such as Indigenous people (Kukutai et al., 2016) to geopolitics of AI (Larssen, 2022). This paper aims to address an underappreciated dimension: the challenges to digital sovereignty in labor relations. Focusing specifically on workers in Artificial Intelligence development. The paper is guided by the following research question: what are the challenges to the sovereignty of AI data workers from the Global South on cloudwork platforms? Discussing the organization of the labor process on cloudwork platforms through a lense of historical-dialectic materialism, it highlights challenges posed by workers’ structurally weak bargaining power and lack of access to labor rights and social protections, arguing that such factors limit the individual and collective sovereignty of workers. Such challenges are illustrated by discussing two major cloudwork platforms: Amazon Mechanical Turk and Microworkers.

Keywords

• digital sovereignty
• artificial intelligence
• data work
• digital labour platforms
• cloudwork
• ai