Social reproduction of AI: Lived experiences of home-based workers on data annotation platforms


Asmita Bhutani, York University

This paper is part of a broader ethnographic study examining women workers’ conditions of work and social relations in transnational data annotation platforms, focusing on gendered and racialized dynamics in this platform work. The proliferation of the Artificial Intelligence (AI) industry and machine learning companies fundamentally rely upon human labour of data annotation. The potential of platform technologies occupies the heart of debates on digitally mediated work and the disruption of traditional norms about workplaces and employment. The paper focuses on the experiences of women who form a substantial part of the data annotation workforce. Although most data annotation platform companies operate from the US and Europe, the labour force for this low-wage, piece work is primarily home-based workers in Global South countries, and India has emerged as a significant location for this work. As a feminist scholar, I center my attention on women’s lives and experiences, revealing the gendered and racialized dimensions of platform work and challenging typical notions of freedom, inclusion, and an available workforce. The paper draws on Marxist feminist framework and ethnographic methods. Specifically, I present how women navigate their paid and unpaid work across their labouring time within the family and on platforms. Focusing on the role each of these institutions play, I discuss how these reproduce racist and patriarchal relations and impact women’s political and economic positions across paid and unpaid “workspaces”. Presenting a range of data collected from semi-structured interviews and home visits of women workers in different parts of India working as home-based workers, I argue that the family and the platform companies play a key role in reproducing feminized platform work, normalizing intensive working conditions for women and naturalizing racism in the highly globalized AI platform labour market. Platforms actively create time zone hierarchies, racial wage codes, accent racism and menial work for racialized workers rendering them politically vulnerable and under-employed while individualizing the risks and responsibilities to workers themselves. Family setups of these home-based workers on the other hand, normalize unequal power relations, perpetuate the devaluation of their identities, all of which routinize highly controlled and divisive labour. These setups shape the spatial and temporal conditions for the production of capitalist commodities and, in their heterosexual form, force women to perform “life-choking” work (Lewis, 2022). These family setups also reproduce gendered worker subjectivities that reinforce the proliferation of casualized arduous and exploitative working conditions prevalent in precarious on-demand platform work. Overall, these patriarchal and racist social relations shape women’s ideas of themselves as workers, their political subjectivity on the platforms, and their imaginaries of resistance against families and paid work. I conclude my presentation with the discussion of responses from gig union organizers in India regarding the challenges of mobilizing this workforce but also women’s responses who have managed to develop informal social media-based collectives as part of their resistance. In discussing these responses, I offer a critique of the existing state of and concerns around platform organizing from a feminist perspective and the possibilities of expanding the organizing agenda for platform work towards a more collective working-class struggle against capitalist political economy. 

This paper will be presented at the following session: