Exploring AI Ethics Through Intelligent Everyday Objects

Abstract

This paper examines artificial-intelligence ethics through DOOR and BOX, two familiar objects whose behaviour depends on a computer-vision system’s prediction of a participant’s ethnicity. By granting access or dispensing a reward only when the system classifies the participant as white, the artworks turn opaque algorithmic discrimination into a direct embodied encounter. The paper discusses bias, classification, power, and the political assumptions embedded in technical systems.