Research by AI Design: The Diversity of Spatial Cycles

ARID LOGICS: The Diversity of Spatial Cycles was a one-week workshop I taught as guest instructor for the ARID LOGICS master’s thesis unit at Taubman College of Architecture and Urban Planning, University of Michigan, in February 2025, coordinated by Kathy Velikov. ARID LOGICS investigates architecture as a response to the environmental and infrastructural challenges of the Mojave’s Las Vegas Valley, a region facing extreme heat, water scarcity and ecological degradation. The workshop results were shown in the exhibition Research by AI Design: The Diversity of Spatial Cycles at Taubman College from February 10 to 17, 2025, which opened with my public lecture Humanizing AI Technology Through Circular Systems: Bridging Design, Ecology, and AI on February 10.

The workshop explored the potential of AI in architectural research through circular design strategies and tectonic exploration. Each student brought their own thesis research; over four steps they visualized spatial cycles, drew programmatic diagrams and experimented with AI-generated visual speculations that connect environmental, technological and communal themes with architecture. An introductory lecture, AI and Visual Tectonics, set out AI as a research tool and its links to historical design thinking. Mixing Realities used text-to-image workflows to fuse non-human elements, natural systems and technology with human spatial experience: public space, water infrastructure, cooling strategies and material behaviour in the desert. Programmatic Diagramming with AI translated the individual research topics into axonometric diagrams that ground conceptual explorations in programmatic and spatial relationships across scales. A final Tectonics session bridged these diagrams with AI-generated tectonic studies, tested as speculative design frameworks and refined in one-on-one sessions.

The workshop introduced AI as both a creative and an analytical tool. No prior AI skills were required: participants worked with publicly accessible platforms, ChatGPT and Midjourney, through step-by-step guidance, learning to use generative AI as a design partner, to articulate research ideas as tectonic expressions and to translate research propositions into architectural form across scales. By bridging architectural tectonics, ecological dynamics and AI-driven research, the students produced design speculations on material intelligence and climate-adaptive urbanism for the Las Vegas Valley, thesis directions they later carried into their individual projects, from a resilience hub to yucca farming for fruit and fibre.