Men and women across India are paid to film themselves doing menial jobs while developers train artificial intelligence robots to perform household tasks.
Transcription
SCOTT DETROW, HOST:
Earlier this month, NPR international correspondent Diaa Hadid went to a small town in southern India called Karur. He reported on how factory workers are being forced to train the artificial intelligence robots that would replace them. It also found a thriving industry for training robots to do menial jobs.
(SOUNDBITE OF CHOPPING CARROTS)
DIAA HADID, BYLINE: Twenty-four-year-old Sneha Jegadish cuts a carrot in her kitchen. It’s a normal scene in a middle-class house in a small Indian town, except Jegadish has connected a smartphone to headphones. It is recording every movement of your hand.
(SOUNDBITE OF ARCHIVED RECORDING)
SNEHA JEGADISH: I put a flat cloth under the (ph)…
HADID: AI industry experts say there are thousands of people like Jegadish across India, who over the past year have found work recording themselves this way. And for about two or three hours each day, Jegadish films himself folding laundry, mopping the floor, any household chore he can find, basically, and makes about $120 a month.
JEGADISH: Double folding, cloth folding. Also the kitchen washing (inaudible). (Speaks Tamil).
HADID: There is a huge demand for this type of material because artificial intelligence companies need millions of hours of these recordings to feed them into artificial intelligence robots so they can perform household tasks. This data cannot be extracted from the Internet.
Prathosh AP is a professor at the Indian Institute of Science in Bengaluru. Like many South Indians, he uses initials to refer to his last name. He specializes in machine learning systems in data-sparse environments. He says this is called physical AI.
PRATHOSH AP: Think of today’s AIs as disembodied brains, right? They have no agency to act physically because they have no body. Now we have to give them all the motor skills.
HADID: Because they are brains and they need to learn to use their bodies.
(SOUNDBITE OF ARCHIVED RECORDING)
UNIDENTIFIED PERSON: Today we launched Neo (ph).
HADID: The demand for this type of data has created opportunities for Indian companies like Objectways. It occupies some buildings on the main street of Karur.
(SOUND OF CAR HORNING)
HADID: Rajeshkumar Ramaswamy, a senior employee at Objectways, tells us they pay thousands of people to film themselves working.
RAJESHKUMAR RAMASWAMY: Philippines, then Nigeria. All over the world.
HADID: From all over the world. Ramaswamy’s team also does some internal recording, he says, to get better quality data.
(SOUND OF FOOTSTEPS)
HADID: We went up a few floors to a floor set up like a house.
Oh my god. It’s like an uncanny valley, huh?
Like in a television studio.
RAMASWAMY: This is a living room. So that’s the kitchen. Then we have three bedroom configurations here.
HADID: In the set up kitchen, Rani N. films herself cutting onions. Like other South Indians, he uses an initial to refer to his last name.
(SOUNDBITE OF CHOPING ONIONS)
HADID: Then he’ll film himself doing other jobs.
RANI N: Next time, carrots, beans. We change the task like cleaning the table, cleaning the closet.
HADID: Ramaswamy says that at this rate, home robots could be sold in India in the next few years. Think about how a robot could help your own housewife, who is raising her young son.
RAMASWAMY: He’ll be preparing food, doing laundry. My wife will rest and a humanoid robot will do everything.
HADID: The tasks will be performed by a humanoid robot. Back at Sneha Jegadish’s house, the young woman who films herself doing housework, is pretty sure that AI robots doing housework are coming…
JEGADISH: (Speaks Tamil).
HADID: …And he can’t wait for the day when a robot takes over this tedious job, even if it means he loses this job, the one he has now.
Diaa Hadid, NPR News, Karur, India.
(SOUND OF MUSIC)
Copyright © 2026 NPR. All rights reserved. Please visit the terms of use and permissions pages of our website at www.npr.org for more information.
The accuracy and availability of NPR transcripts may vary. The transcript text may be revised to correct errors or match updates to the audio. Audio on npr.org may be edited after its original broadcast or publication. The authorized record of NPR programming is the audio record.
