AI & Labor
The people training the world's AI: a day with Kenya's data labellers
Behind every clever model is a crowd of people tagging data, frame by frame.
At six in the morning, in a Nairobi office block, labellers begin tagging data for the next generation of AI models — outlining objects in images, judging whether a passage of text contains violent content.
It is invisible work. The smarter a model looks, the more it seems to have "taught itself" — yet every capability rests on thousands of hours of human labelling.
Hourly pay, psychological strain, bargaining power — these questions are being seen by more and more people.
About this story
Compiled and interpreted by AI闹.
Original:Sam Otieno · Rest of World
The Workers Behind the World's AI →Editor's note: A few original terms are kept to preserve the industry context.
