When Generative AI Starts With Architectural Intelligence, Not Prompts

Architectural designer and AI researcher Samuvel Benhursha proposes a constraint-first framework where generative AI follows architectural reasoning instead of prompts. Backed by a 2025 IEEE paper, GAN-based modeling, and measurable metrics like PSNR and SSIM, his method encodes cultural, climatic, and structural logic before generation. Recognized through lectures, competitions, and exhibitions, the approach positions AI as accountable architectural intelligence.