The dominant image of AI and work is that of a Silicon Valley programmer using ChatGPT to write code faster. But this image conceals a much more complex and unequal reality. To understand who benefits and who loses from AI-based automation, we must look at the entire value chain — from the cobalt mine to the data centre, from the data labeller to the CEO.
At the base of the chain are workers who extract raw materials: cobalt for batteries (often mined by child labourers in the Democratic Republic of Congo), lithium for chips, rare earths for electronic components. These workers do not benefit from AI — they pay its price with their health and their environment.
One step up are those whom Mary Gray and Siddharth Suri call "ghost workers". They are the hundreds of thousands of people in the Global South who label images, transcribe audio, moderate violent content and evaluate AI output quality for platforms like Amazon Mechanical Turk, Scale AI, Sama. They earn a few cents per task, have no stable contracts, no union protections, no benefits. They are the invisible workforce that enables AI to function.
In 2023, a TIME investigation revealed that OpenAI had paid Kenyan workers less than 2 dollars an hour to moderate the violent and sexual content used to train ChatGPT, exposing them to traumatising material without adequate psychological support. This is the human cost that does not appear in presentations about the wonders of generative AI.
At the top of the chain are the major corporations — Google, Microsoft, Meta, OpenAI, Amazon — which capture the vast majority of value generated by AI. These companies have a combined market capitalisation of trillions of dollars but employ relatively few people. Productivity increases, profits increase, but benefits concentrate in an ever-smaller number of hands.
In between are workers in wealthy countries, whose fate varies enormously. Highly skilled professionals (programmers, doctors, lawyers) tend to use AI as a productivity amplification tool — at least for now. Workers in routine roles (accounting, customer service, translation) see their tasks automated or devolved to hybrid systems where AI does most of the work and the human handles exceptions. Gig workers (delivery riders, drivers, waiters) are increasingly managed by algorithms that dictate their rhythms, routes and compensation.