Structure and Ecology – „CARBON INTELLIGENCE: AI-Assisted Development of Structural Sustainability“

Every structure has an ecological cost, and it is spread very unevenly across a building. Some elements carry a lot of load with very little material. Others cost a great deal of carbon and do very little work. Which is hard to see in a design, because the information sits inside engineering models that architects usually meet only as finished drawings. The seminar opens those models up and asks a simple question: how much carbon does it cost to carry a load down through this building, and where is that cost worth it?
Students answer it by building their own applications, using vibecoding tools that turn a description in ordinary language into a working software. No programming experience is required. Learning to direct and judge AI-generated code is the real skill of the seminar, and it carries over to anything you want to build in the future.

The work happens on real structural models from the engineering office Bollinger + Grohmann. Students first build a small tool that gets the data out of those models, then an interactive dashboard on top of it: a building you can turn, filter and colour by ecological cost, so the expensive parts of a structure become visible in the geometry. Everyone leaves with an application they built themselves, a sense of which design decisions shape a structure’s footprint, and a healthy scepticism towards the measure they used to judge it.