AI POWERED IDEATION

Masters Dissertation

Glasgow, UK / 2024

Description:

The Masters dissertation studies artificial intelligence, already used in architecture, and investigates AI’s possible impact on it. The theoretical part explores the structure, operation, and training methods of different AI models and explains how they evolved from optimisation tools and the “black box” problem behind them. This part also reviews the position of architectural researchers and data scientists assigned to AI within the digitalisation and automation of the architectural industry. The practical part investigates how “creative” AI tools can be integrated into the architectural design ideation process using my Stage 5 Final Design Thesis as the subject of study and training dataset. The finished design media (drawings, diagrams, visualisations and physical models) were translated into text prompts using the semantic method to be “recreated” by TextToImage AI. Then, a brand-new design iteration on the same site was developed from scratch using TextToText, TextToImage models and AI-powered CAD. The results of the practical part demonstrated which architectural design tasks AI can handle already and possibly in the future. This, together with the theoretical research outcomes, opened the route for the final discussion of AI’s possible impact on the architectural design process, employees’ skill sets, supply chain, legislation, and ethics behind it.