The recent surge in commercially available AI for text and image generation has made this technology available to architectural practitioners and scholars, who increasingly use it to support different design stages. This chapter visually interprets one of the most widely used AI models—artificial neural networks—to reveal the mechanisms that enable machines to classify, predict, and generate new data, including novel design propositions. Fundamental concepts such as dataset, training, and function approximation are explained through three-dimensional visualizations and graphs, which can be recreated in Grasshopper, the visual scripting environment of the CAD software Rhinoceros 3D. The chapter also explores the relationship between AI and computational design methods, clarifying AI’s unique characteristics and offering insights into its integration within existing design workflows. By uncovering the computational and mathematical logic of AI, this chapter provides readers with a deeper understanding of its inner workings and evolving role in architectural design.