The gig workers who are training humanoid robots at home

AI Summary

According to MIT Technology Review, a growing number of gig workers across more than 50 countries are being hired by Palo Alto-based company Micro1 to record themselves performing household chores using iPhones strapped to their heads, with the footage sold to humanoid robotics companies as training data. Workers like 'Zeus,' a medical student in Nigeria, earn $15 per hour — competitive pay in local economies with high unemployment — recording tasks such as folding laundry, washing dishes, and ironing clothes. Micro1 CEO Ali Ansari estimates that robotics companies are now spending more than $100 million annually on real-world movement data, with investors pouring over $6 billion into humanoid robotics in 2025 alone. Competitors including Scale AI, Encord, and DoorDash are running similar data-collection operations, while China has established state-owned robot training centers where workers use VR headsets and exoskeletons. Scale AI has reportedly gathered more than 100,000 hours of footage to date, though UC Berkeley roboticist Ken Goldberg cautions that training humanoid robots may ultimately require data volumes comparable to the text corpora used for large language models — which would take a human 100,000 years to read. The article raises significant concerns around informed consent, data privacy, and quality control, as workers are not told which robotics companies receive their data or how it is stored and shared with third parties.