ACADEMIC
RE-DESIGN
GENERATIVE AI IN ARCHITECTURE
TOOLS
Stable Diffusion XL, Google Colab, Rhino 7, Grasshopper, ChatGPT, Gemini, PromeAI, AutoTRIZ, Photoshop, IllustratorCOMPLETED
October 2024
This project explores a Generative-AI driven redesign of Sankofa Square, formerly Yonge &
Dundas Square, in Toronto, Ontario. The re-designed public square addresses urban “dead
spaces”, which are defined here as spaces that remain empty and underutilized during times
of non-organized programming. This can lead to the area becoming unsafe and uninviting,
overall contributing negatively to the urban environment. Design objectives are achieved
by combining context-specific ideation using text-based generative AI models, with agentsimulation of pedestrians using a grasshopper script. Design interventions are formulated
using AI, which are then applied to the site and visualized using agent-simulation as context.
︎
DESIGN OBJECTIVES
INTERVENTIONS
- Interactive Pathways, Kinetic Tiles- Playful Pavement
- Sound & Temperature Responsive Features
AGENT SIMULATION
- Movement Simulation- Path & Engagement Analysis
- Layout & Configuration Optimization
BENEFITS/IMPACT
- Encouraging Movement & Exploration- Foster Social Interaction
- Cultivate a Sense of Ownership
AGENT SIMULATION
Grasshopper
Identify Form
Visualize Data
FINAL VISUALS
“HERO” IMAGE, ARIAL VIEW
POV: WALKING THROUGH
POV: COMING UP THE ELEVATOR
PLAN VIEW: NIGHT-TIME
PLAN VIEW: PARTI
