Sorting the Harvest at Scale: A Governance Framework for Generative AI, from Brand Systems to Urban Futures

Preprint. Intended for submission to a peer-reviewed journal.

Purpose. Generative AI's migration into urban design inherits, at much larger scale, an unresolved practitioner problem: how architects and planners evaluate, govern, and responsibly adopt AI tools before applying them to work of consequence. This paper asks whether a governance framework built at building and brand scale extends to the scale of generative urbanism.

Methods. Practice-based research presents AI-KORP, a practitioner framework ("Basket & Ripeness") for auditing generative AI tools across brand, spatial, and process work. It tests it against two bodies of evidence: a proof-of-concept classification pipeline — AI-KORP | Neural Asset Archives — run first against a single building and then extended to aerial imagery of six logistics facilities at urban scale, yielding a three-type brand-architecture fusion typology; and the branded logistics landscape of Germany itself, in which national retail and delivery brands apply one colour and mark system across facade, fleet, and signage.

Results. Brand and building recognition prove to be a single, fused visual layer rather than separate systems at both the building and city scales. An aerial survey of six German logistics facilities yields three distinct fusion types: fleet-carried signal with a neutral envelope (Type A), envelope-carried signal readable from altitude without the fleet present (Type B), and hybrid cases in which the envelope and the fleet carry the brand signal simultaneously across all layers (Type C). In all three types, the fusion precedes the classifier — it exists in the built environment before any AI system touches the imagery.

Conclusions. Any credible framework for generative urbanism must specify not only what AI can generate, but also the ripeness, sovereignty, and disclosure conditions that must be met before generative or classification outputs are treated as design or planning authority — a governance-first, rather than capability-first, approach to scaling generative intelligence from the object to the city.

Full methodology, code, and generated figures: https://github.com/archi-netizen/ai-korp-neural-archives