The term “artificial intelligence” is misleading. The intelligence on display in modern chatbots is not entirely artificial; a significant portion of it is appropriated from the collective knowledge of a vast, underpaid human workforce. Tech companies are, in effect, mining the intelligence of thousands of people and rebranding it as their own technological creation.
This process of appropriation happens in the daily work of AI trainers. These individuals, often with specialized expertise in fields like science or the arts, are paid to pour their knowledge into the machine. They correct the AI’s mistakes, rewrite its awkward phrasing, and teach it the nuances of their specific domains. Their intellectual property and expertise are being systematically transferred to the model.
However, these human contributors receive no ownership or long-term stake in the powerful technology they are helping to build. They are paid a low hourly wage for their time, while the immense value they create is captured entirely by the tech corporation. One researcher aptly described this dynamic as a “pyramid scheme of human labor,” where value flows upwards from a broad base of expendable workers.
This raises profound ethical questions about the nature of AI. When a chatbot provides a brilliant summary of a complex topic, are we witnessing an act of machine intelligence, or are we simply reading the polished work of a hidden, under-compensated human expert? The industry’s success is built on blurring this line, presenting appropriated human intelligence as a purely artificial marvel.