Meet India’s newest content creators – they work for robots

MUMBAI: Nagireddy Sriramyachandra straps a smartphone to her forehead every morning and gets to work. She slices mangoes, folds laundry, makes coffee. If her hands drift out of frame, the app barks at her: hands not detected. She shoots up to 90 four-minute clips a day and earns Rs 250 an hour ($3). Nobody watches a single second of it.

“Who else will give you 250 rupees an hour just for doing housework?” she asked, with impeccable logic.

Sriramyachandra, 25, is not a content creator chasing likes or a YouTuber angling for brand deals. Her footage will never air on a television network. Her audience is entirely algorithmic. She is part of a fast-growing, largely unheralded gig workforce producing what the industry calls egocentric data: first-person spatial video captured from the human point of view. Global AI data firms, Objectways among them, are buying this footage by the terabyte. Processed through platforms such as Amazon SageMaker, it trains the next generation of humanoid robots being built for Fortune 500 multinationals. The woman doing the laundry is teaching the robot how to do the laundry.

For India’s media and broadcasting ecosystem, that sentence should land like a thunderclap.

The operation extends well beyond home-based gig workers. Objectways runs dedicated physical studios in industrial hubs such as Karur in Tamil Nadu, where young engineering graduates — among them 21-year-old Rani N. — don smart glasses and motion-capture gear to fold the same towel up to 90 times a day on a mock bed, in every conceivable position. Once a few thousand hours of footage are in the can, the studio swaps the wallpaper and bedding to give the client’s neural networks the visual variety they need to generalise. Then the whole cycle begins again.

The executives running these firms are sanguine about what they are building. Ravi Shankar, head of Objectways, put the corporate view plainly: “Some jobs are supposed to be taken over, so humans can go and do better things.”

On a roadside in Bengaluru, Ponni sees it rather differently. The 55-year-old has spent a decade making flower garlands. She too was recently paid to sit with a phone strapped to her forehead, mapping her hand movements for an AI model. Her assessment of the technology carries none of Shankar’s breezy optimism. “The next generation who might have to do work similar to mine,” she said quietly, “they will face a problem.”

That tension, between the immediate financial lifeline and the long-term displacement it finances, sits at the moral core of one of the most consequential labour stories in India today. The people being paid to produce this data are not tech workers with options. They are homemakers, textile hands, garland makers and engineering graduates in tier-two towns, for whom Rs 250 an hour represents genuine economic flexibility that traditional broadcast work, local retail and call centres have never been able to provide at scale. They are not unaware of the bargain they are striking. They are simply making the most rational choice available to them.

From a pure business perspective, the structural logic of the sector is startling. Traditional production houses pour millions into pilots and promotional campaigns, shouldering enormous financial risk in the hope that a show survives its first season. The spatial data production sector operates on an entirely different model: pure B2B contracts, signed before a single frame is shot, with a client locked in from day one. Demand for high-fidelity situational footage is insatiable. Production costs are anchored to micro-wages. There are no ratings anxieties, no advertiser negotiations, no subscriber churn to manage.

India has spent decades positioning itself as the world’s back-office for VFX, animation and post-production, adding gloss to Hollywood blockbusters and domestic streaming giants alike. The spatial AI boom is building a parallel track that is potentially far more lucrative and far less glamorous, and it is scaling rapidly.

Yet the macro policy conversation has barely begun. NITI Aayog has warned explicitly that while global and domestic debates about AI focus almost exclusively on how large language models will disrupt white-collar professionals — coders, writers, legal assistants — almost no attention is being paid to how physical automation will affect India’s 490 million informal workers, the precise people now being recruited to train the machines. India is marketing itself aggressively as the world’s premier engine for AI data processing while exporting the raw intelligence of its working class to fuel foreign intellectual property.

Entrepreneur Ramesh Srivats made the strategic point sharply on social media: why sell this data cheaply when an Indian company could use it to build its own AI-powered domestic robot? He suggested calling it B.AI, a play on bai, the colloquial term for domestic help. It was a joke. It was also a serious question that nobody in government or industry has yet answered.

The definition of a video asset is changing permanently. The pixels moving across Indian screens are no longer purely for human entertainment. They are corporate assets, training data, the raw material of a robotics revolution. Media executives who spend their days tracking TAM AdEx volumes and optimising prime-time slots need to understand that an entirely new content economy has taken shape around them, one with zero audience, guaranteed revenue and a client list that makes any broadcaster’s advertiser roster look modest.

Sriramyachandra, for her part, remains cheerful. “I may get a robot myself in the future,” she said.

She may well be right. The more uncomfortable question is whether that robot will have learned everything it knows from watching her.

VIDEO SOURCE: AFP NEWS AGENCY