Project by Chelsea Hai @chelseahai_
HAI is a garment intelligence system developed at Columbia GSAPP that redefines fashion as an algorithmic response to personal narrative and environmental data. The system bypasses traditional industry hierarchies by translating a user’s daily context, destinations, transit modes, and aesthetic intent, into optimized, 3D-printable structures. The technical pipeline utilizes a dual-LLM extraction process to geocode travel routes and analyze contextual nuances. By cross-referencing this data with real-time weather and routing APIs, HAI computes eight specific metrics, like Fit, Mesh, and Airflow, through logic-based formulas. These values are streamed directly into Grasshopper to generate bespoke G-code for TPU fabrication. Moving beyond mass standardization, this system proposes a shift toward distributed, on-demand manufacturing. By utilizing recyclable TPU, the project establishes a closed-loop system where garments are treated as temporary “"data records”“ of a specific moment. When a garment’s utility ends, the material is shredded and recycled, providing a sustainable, regenerative alternative to the linear fashion economy.
