Dear Artificial Intelligence, how do I write about architecture? (HS 23)

Seminar Architectural Criticism (052-0815-23), Antje Stahl, Fridays, 11:45-13:30, HCP E 47.1 / HIL F 10.3

Enlarged view: Robot looking at Swiss Architecture, generated by Deep AI — Robot looking at Swiss Architecture, generated by Deep AI

Enlarged view: Swiss Architecture, generated by Deep AI — Swiss Architecture, generated by Deep AI

If you ask ChatGPT these days how to write about architecture, the OpenAI program gives some advice: “first you have to understand architecture, read books and magazines, visit and sketch different buildings for inspiration. Only then should you move on to description and analysis, preferably emphasizing what is so special. And last but not least, connect everything to society.”

Sounds pretty good. The question is, how can ChatGPT and other AI writing programs help us do just that? This fall semester's seminar on architectural criticism focuses on text production using the new AI tools.

Enlarged view: Als OpenAI und andere grosse Sprachmodelle lanciert wurden, versetzten sie Hochschulen ja eher in Angst und Schrecken. Es bestand die mitunter berechtigte Sorge, dass wissenschaftliches Arbeiten an Maschinen delegiert und kritische Reflexion damit verlernt würden. Diesem Vorurteil soll in einer Reihe von Übungen entgegengewirkt werden, in denen die SeminarteilnehmerInnen in den Dialog mit der Maschine treten sowie eigene Schreibfähigkeiten und das Redigat erlernen. Im Zentrum der gemeinsamen Textkritik stehen diskursive Fragen zu Konzepten der Intertextualität und Autorschaft, Kreativität, Deepfakes und KI-Ethik. Bilder: Deep AI generiert — Architect looking at Swiss Architecture, generated by Deep AI

When OpenAI and other large language models were launched, they tended to scare universities. There was sometimes a justifiable concern that scientific work would be delegated to machines and that critical reflection would be neglected.

This prejudice will be counterd through a series of exercises. Participants of the seminar will enter into a dialog with the machine and improve their own writing and editing skills. Textual criticism will revolve around discursive questions on concepts of intertextuality and authorship, creativity, deepfakes and AI ethics.

Pictures: Deep AI generiert