The rise of artificial intelligence (AI) and cloud-based processing has created a wealth of opportunities across all industries. It offers us the power to approach complex computing problems with incredible speed and accuracy that exceed our human capacity. What would have taken humans weeks to complete can now be resolved in hours. This is the basis for a revolution in the design process which can change our approach to planning buildings and cities.

To explain generative design best, let’s have a look at nature. When we consider a flower, for example, we marvel at its perfection. Every nuance is perfectly calibrated to capture moisture, to survive in harsh conditions, and to reproduce in a living ecosystem. However, this flower did not come to be overnight. It took millions of years of evolution. Millions of design iterations one after the other, trial and error, update after update. This flower collected information from its existence to change its properties over time until it reached the peak of performance and efficiency. Generative design emulates this process of evolution.

Generative design creates a virtual world to test versions of a design solution. Various data are entered, for example about building use, building height, construction area, loss rate, cost of construction, GHG emissions targets, number of residents, availability of green space, access to natural light, connectivity, and proximity to transport. A generative design process then renders this information into a series (hundreds) of typologies and spatial development options through computational design: the basic building blocks to inform detailed master planning. However, it doesn’t stop there. Each typology is tested and validated based on simulating various conditions, comparing one typology against the other to identify weaknesses and redundancies. This creates stronger and stronger design options the model achieves acceptable operational performance based on project success criteria. In essence, thousands of hours of design evolution are being condensed into a single afternoon – based on the power of the processor that is running the computations. It would be impossible to achieve the same output through traditional, human-powered processes.

This does not mean that we should replace designers. Instead, we should use computational design to provide a defensible and quantifiable spatial development framework to guide detailed human-led design. Instead of doing feasibility analysis and concept design, human designers are able to focus on creative decision-making associated to experience design, while relying on data-driven feasibility and project resilience.