Psychosocial and Health-Related Risks in AI value chains: the Case of Latin American Data Workers

This article examines the psychosocial risks and mental health implications of data work in Latin America through the Job Demands–Resources (JD–R) theory. While research on global AI value chains has documented economic, social, and political disadvantages in the Global South, far less attention has been paid to psychosocial and health-related risks affecting workers essential to the global development of AI. Our analysis draws on a survey of about 400 data workers in Latin America (Brazil, Mexico, Chile, Venezuela), recruited through two microwork platforms, Prolific and Microworkers. In addition to sociodemographic variables, we employed the Spanish Burnout Inventory to assess burnout. Preliminary findings show 38% of respondents report high psychological exhaustion and 27% low work-related enthusiasm—clear signs of fatigue and disengagement. Indolence (17%) and guilt (5%) appear at lower levels, suggesting the most severe forms of burnout are rare, though vulnerability remains significant. The data show that high demands (monotony, tight deadlines, income insecurity, algorithmic surveillance, social isolation) converge with scarce resources (limited autonomy, lack of variety, absence of social protection). Workers with longer exposure to platforms report greater exhaustion and indolence; those relying exclusively on microwork show higher exhaustion and guilt; and those perceiving a mismatch between education and tasks report lower enthusiasm and greater indolence. Within the JD–R framework, we argue that AI value chains not only deepen structural disadvantages but also generate psychosocial risks that reinforce them, and must be integrated into debates on political economy of platforms and AI in the Global South.