In the current architectural imagination, “Geo/spatial Intelligence” has become the mot du jour–a term often deployed by emerging tech platforms to describe the transformative power of Large Language Models (LLMs) and their world-building generative capabilities. Yet, for designers, the intelligence of space and geography is an age-old inquiry. Every era of design is marked by the aesthetic and spatial trace of its tools; if the LLM is our tool du jour, then this course interrogates the specific architectural trace it leaves behind.

This class will move through three interwoven modules: 
1. Tools + Reasoning introduces students to a suite of tools for AI-enabled spatial production.
2. Designing Hallucinations investigates the generative glitch not as a failure, but as a critical site for design.
3. Geospatial definitions will step back from the current AI rush to revisit the foundational texts and design practices that predate neural networks and generative AI but confront the same core concepts.

Students will use their studio projects as departure points to become familiar with the tools for iterating with AI. Broadly, the sequencing of the three modules should align with that of a semester-long design project. Module 1 will focus on site research and geospatial data as the language of design. Module 2 will use in-progress, unfinished designs as seed for hallucinatory resolutions. Module 3 will conclude with a manifesto based on a study of texts and on the semester’s learnings. Note that it is not required that students take a semester-long design studio in conjunction. Please reach out to the instructor to discuss alternative project approaches.

The final deliverable for the semester will be a collective contribution to a volume–a manual of experiments, drawings, and texts that seeks to provide rigorous, designer-led definitions of and positions on geo/spatial Intelligence.