“The good ones go into the pot, the bad ones go into your crop.” – that’s how the story goes in Cinderella. But who actually decides what’s good and bad? And what happens when complex decision-making processes are increasingly automated and delegated to “intelligent” systems?

Upon entering the exhibition space, visitors encounter an industrial-looking sorting plant. Using a feeding mechanism and a conveyor belt, the “EZ Quality Sorter V2” automatically separates, analyzes, and sorts pea seeds into good and poor quality. If a pea is categorized as “bad,” it is sent to the reject bin. If it is categorized as “good,” it eventually enters the container for further processing.
Visitors that approach the workstation are invited to take a seat and visually inspect the peas one by one through an optical device. They are asked to assess their quality and enter their selection by pressing either a green or red button. With each button press, the machinery takes a close-up picture of the respective pea and, according to the user’s selection, adds it to an image database. Once the user leaves the station, the machine automatically continues the sorting process based on the previous user inputs.

Today’s “intelligent” systems often run on invisible human labor and subjective decision-making processes that have been crystallized into hard facts through formalization and automation. When the machine appears autonomous, one easily assumes that its decisions are neutral, objective, and rational. However, when one becomes the decision maker, one quickly finds oneself on shaky ground. Classifying a complex world and reducing it to pregiven binaries turns out to be a vague, troubling, and even violent endeavor.