Text2Forest Part 2 Project 2025 Xavier Thanki The Bartlett School of Architecture, UCL | UK Text2Forest explores language-based machine learning as a generative design tool, positioning artificial intelligence not as a means of optimisation but as a co-author of architectural form. It responds to the urgent need to reprogram how landscapes are understood and shaped - challenging forestry practices that continue to prioritise monoculture, speed, and uniformity over ecological intelligence. In the face of accelerating climate change, this project proposes a more adaptive, responsive approach, where data, environment, and identity are intertwined.At the core is a fine-tuned language model that translates weather-based text prompts into tree geometries, each probabilistically generated from environmental inputs. The tool, Text2Forest, is deployed across a site to algorithmically plant a forest: each tree unique and tied to its own climate narrative. This forest is not decorative, but productive: a living index of environmental variation and a material resource for design.By merging natural language processing with procedural growth algorithms, the project reframes both forestry and architecture as acts of co-creation. It embraces indeterminacy and decentralised authorship, embedding ecological logic directly into form-making. Architecture no longer begins with an idealised object, but with the generative conditions of the forest itself. Tutor(s)Ms Abigail AshtonMr Andrew Porter