Investors are pouring money into solving this challenge, spending more. 6 billion dollars On humanoid robots in 2025. And data recording at home is becoming a fast-growing economy around the world. Data companies like Scale AI and Encord are recruiting their own armies of data recorders, while DoorDash Pays delivery drivers. To film your actions. And in China, workers at dozens of government robot training centers wear virtual reality headsets and exoskeletons to teach humanoid robots how to open a microwave and clean a table.
“There’s a lot of demand, and it’s growing really fast,” says Ali Ansari, CEO of Micro1. He estimates that robotics companies are now spending more than $100 million a year to buy real-world data from his company and others like it.
A day in the life
Workers at Micro 1 are screened by an AI agent called Zara that conducts interviews and reviews samples of work videos. Each week, they submit videos of themselves working around their homes, following a list of instructions about things like keeping their hands exposed and moving at a natural pace. Videos are reviewed by both AI and humans and either accepted or rejected. They are then interpreted by AI and a team of hundreds of humans who label the actions in the footage.
“There’s a lot of demand, and it’s growing really fast.”
Ali Ansari, CEO of Micro1
Because this method of training robots is in its infancy, it is not yet clear what makes good training data. Still, “you need to give the robot a lot of flexibility to generalize the basic navigation and manipulation of the world well,” says Ansari.
But many workers say it’s a challenge to create a variety of “chores” in their tiny homes. Zeus, a scruffy student living in a humble studio, struggles to record anything other than ironing his clothes every day. Arjun, a tutor in Delhi, India, takes an hour to make a 15-minute video because he spends so much time doing new tasks.
“How much material (can be made) at home? How much material?” He says
There is also the sticky question of privacy. Micro1 asks workers not to show their faces on camera or reveal personal information such as name, phone number and date of birth. It then uses AI and human reviewers to remove anything.
But even without faces, the videos capture a deeper slice of workers’ lives: the interiors of their homes, their possessions, their routines. and understanding what personal information they may be recording when they are on camera. Such footage reviews may not filter out more sensitive information than the most obvious identifiers.