This is an agenda-setting paper from an ongoing multi-country project that establishes a conceptual foundation for analysing AI-related microwork, outlining a new framework for understanding how inequality is structured across global AI value chains. Its central aim is to investigate how intra-national stratifications—class, caste, gender, and race—interact with transnational hierarchies to structure AI microwork, and how these intersecting dynamics shape working conditions, economic dependence, and wellbeing outcomes. Data labellers (ghostworkers or microworkers) perform essential yet largely invisible tasks, including labelling and verifying datasets used to train AI systems. Existing research has demonstrated how low-paid data labelling and annotation work is disproportionately outsourced to workers in the ‘Global South', while higher-value tasks remain concentrated in the ‘Global North' (Graham & Ferrari, 2022; Rani & Dhir, 2024). While this scholarship has been crucial in foregrounding the racialised and colonial dimensions of digital labour (Lopez & Sanchez, 2024), it relies on a spatial binary (global south versus north) that obscures a critical analytical gap: the role of intra-national stratification. This paper addresses that gap by showing how class, caste, gender, and race within countries co-constitute the organisation, distribution, and lived experience of microwork. In doing so, the paper asks whether AI systems are not only shaped by, but actively reproduce, hierarchies of race, caste, class, and gender. This reframing positions microwork as a central infrastructural node in global AI value chains rather than a peripheral activity, highlighting its constitutive role in sustaining AI production. The paper makes three interrelated contributions. First, it develops a theoretical account of how intra-national stratification structures labour segmentation and economic dependence on microwork, moving beyond aggregate accounts of platform labour to show how inequality is embedded within national contexts. Second, the paper shows how intra-national stratification interacts with transnational hierarchies to structure labour within global AI value chains. Rather than treating the Global North and Global South as internally homogeneous, it highlights how layered and relational inequalities are co-produced across and within countries. Third, it analyses how labour and technological infrastructures mediate these inequalities, foregrounding the infrastructural dimensions of inequality and their differentiated impacts across contexts. References 1. Graham, M., & Ferrari, F. (Eds.). (2022). Digital Work in the Planetary Market. MIT Press. 2. Rani, U. & Dhir, R. K. (2024). AI-enabled business model and human-in-the-loop (deceptive AI): implications for labor. In Edward Elgar Publishing eBooks (pp. 47–75). 3. López, I. C. & Sánchez, V. (2024). Artificial Intelligence, Microwork, and the Racial Politics of Care. In M. L. Jones & A. Levendowski (Eds.), Feminist Cyberlaw (1st ed., pp. 165–178). University of California Press.
Subject :
:
Abstract
Topics
:
Transparency and traceability in the AI models and their supply chains
Topics
:
Working conditions, occupational safety and health of workers in the human-in-the-loop
Topics
:
New frameworks for a “human-centric” AI supply chain
Topics
:
Algorithmic management, labour control, and workers’ resistance
Topics
:
Gender and digital labour
Keywords
:
microwork ; AI ; ghostwork ; race ; caste ; gender ; value chain ; worker ; labour
