India is becoming a global hub for AI data work, with workers in kitchens, factories, and studios recording everyday activities to generate egocentric data used to train AI systems.

Contractors like Objectways count major clients such as Amazon SageMaker and serve Fortune 500 firms, positioning India as a key intermediary for AI data collection, processing, and annotation.

Experts expect a future of human–robot collaboration rather than full replacement, with humans supervising and managing robotic systems.

In Tamil Nadu and Andhra Pradesh, factories and labs use head-mounted cameras to capture workers’ movements during tasks like packing and labeling, raising privacy concerns and prompting workers to set self-imposed boundaries.

Ponni, a Bengaluru street vendor who records her work with a head-mounted device, highlights worries that future generations may face similar livelihood challenges as automation expands.

Ethical and privacy concerns persist, with workers like Nagireddy Sriramyachandra setting limits on what parts of home life get recorded, showing tension between data for AI and personal privacy.

Young contributors such as 21-year-old Rani N. in Andhra Pradesh describe the work as tolerable yet intrusive, feeling constantly watched while filming household tasks and conversations.

Across Tamil Nadu and Bengaluru, workers describe repetitive, camera-wearing tasks and a sense of surveillance, labeling the job as tolerable but intrusive.

Experts warn about the broader labor impact of AI, stressing the need to address how automation affects India’s vast informal economy and protect livelihoods.

NITI Aayog cautions that automation could replace jobs and urges proactive measures to help India’s 490 million informal workers benefit from AI while safeguarding incomes.

Think tanks note that AI poses risks for the informal sector even as automation promises new opportunities, underscoring the need for protective policies.

The data-focused approach hinges on egocentric footage from head-mounted cameras, showing how movements are translated into machine-language instructions.