Jiayuan Yu

Artificial intelligence has given rise to a new wave of workers, but compared to the Industrial Revolution era, this working class is almost invisible, akin to the existence of ghosts. Ghost workers enhance machine learning algorithms by performing repetitive tasks online such as image annotation and content moderation. Many workers have been concealed within the promotion of company automation and full automation, even replacing their identities in human society with a series of numbers. They are a marginalized workforce that is not protected by law or society. This project compiles a set of characteristics of ghost workers through extensive research and provides them to AI for visualizing their avatars and virtual work environments. The characters in green suits resemble movie green screen costumes, allowing them to be easily identified and removed by software. The majority of ghost workers’ tasks involve annotating images gathered from the internet and using them as input to train AI models. In other words, they train models with their own work, creating images as if in a Möbius loop, gradually revealing the hidden parts.