People are now being paid to film themselves performing household chores to help train artificial intelligence for domestic robots.
This trend marks a shift in the gig economy, where human physical movement is being digitized to bridge the gap between AI software and robotic hardware. By recording how humans naturally interact with a home environment, tech companies aim to create robots that can navigate and clean without constant human supervision.
Participants in these programs, including those working with the U.S. startup Shift, record a variety of everyday tasks in their own homes [2]. The footage serves as a dataset for humanoid robots to learn the nuances of domestic labor, such as folding laundry or tidying a room, that are difficult to program manually [1].
The workforce for this data collection is global. Workers in countries such as India and Nigeria are participating in these programs to earn income [1]. In some instances, the pay for recording this chore footage can reach as much as $50 per hour [4].
One participant, a medical student identified as Zeus, has filmed himself performing household tasks to contribute to the AI training process [3]. Other workers use head-mounted cameras to provide a first-person perspective of the chores, ensuring the AI understands the exact visual cues and hand movements required for each task [2].
Tech firms are utilizing this video data to develop models that will eventually power a new generation of home assistants. The goal is to transition from robots that perform single, repetitive tasks to humanoid machines capable of complex, multi-step domestic work [1].
“People are now being paid to film themselves performing household chores to help train artificial intelligence for domestic robots.”
This development highlights the growing dependence of the AI industry on 'human-in-the-loop' data collection. By outsourcing the labor of recording chores to gig workers in developing economies and the U.S., tech companies are essentially crowdsourcing the physical blueprints of human domesticity to accelerate the commercialization of humanoid robotics.
