Empty Network explores the intersection of Zen Buddhist metaphysics and the inherent opacity of artificial intelligence. The visual component features a 3D model of the “Iron Seated Buddha” from the National Museum of Korea. The audio and visuals are manipulated through control signals derived from machine learning analysis of the Heart Sutra, where the text is converted into numerical embeddings. This technical process was developed in collaboration with machine learning engineer Sohyun Shin. While traditional sonification aims to make data perceivable, this work operationalizes the “black box” problem of deep learning. By treating these inherently opaque text embeddings as a source of aesthetic noise rather than extractable meaning, the piece subverts the algorithm's intended function and turns its lack of interpretability into a feature.