Based on the artist’s experience of practicing “A Narrative on Calligraphy” for many years, Calligrapedia: A Universal Algorithm #1 utilizes a machine learning model called Generative Adversarial Network (GAN) to analyze a large number of calligraphy drafts. Drawing inspiration from natural elements such as landscapes, animals, and plants, it has developed a series of human-machine-made Chinese characters that flow between script and organic things. On the other hand, through collaboration with a folk tale narrator, it reinterprets ancient oral stories passed down worldwide with “qualitative transformation” as its core. Deconstructing machine learning’s aspects of image style recognition and simulation, the work explores how words, language, and mythological narratives in human history continuously interpret themselves through learning from the world around them. The work constructs an intertwined view of human-machine-nature interaction. Combining the interdisciplinary practices of calligraphy, generative images, and sound performance, it contemplates on the essence of creation and relationships between humans, machines, and nature.