Architecture after artificial intelligence

Excel did not finish off accountants. Photoshop did not kill photographers. AI will not take your job…

— PlatziConf 2023

Artificial intelligence (AI) is reaching society at an accelerating pace, and it is bringing with it advances in the way we visualize reality. One example is Generative Artificial Intelligence (GAI).

This model of AI has been evident since April 13, 2022, when OpenAI published a paper describing in detail what diffusion models are, and made its Application Programming Interface (API) available — allowing an ordinary user to generate an image from text, what is known as text to image.

The evolution of this technology, and others in the same industry, has followed a growth trend unseen in other assets, and it will also optimize the architecture industry: starting from a model trained on images of buildings with a specific desired concept, we can create images or renders in a matter of seconds, depending only on the computing power of the machine used.

Creating these renders almost instantly opens the way to both creativity and efficiency, since creator and client alike can digitize what is in their minds in very little time, without waiting days, weeks or even months to arrive at a satisfactory 3D model. Three-dimensional modeling is skipped, moving directly to rendering, and the work becomes far more directed toward what we are actually looking for — without losing time in digressions and iterations that often cost time, effort, money and clients.

GAI adopts concepts such as the prompt, a textual description of the image to be generated, enriched with punctuation, keywords and further specifications that produce a more precise image of what is expected. The more technical outcome of this practice is what we call prompt engineering.

But the uses of GAI do not stop there: beyond shortening creation times, these tools also stylize models starting even from a simple sketch.

Handling digital instruments of this kind does not replace the designer's creativity; rather, it amplifies and refines it to its most attainable point.

AI is not a substitute for human talent. It is a tool that grows more accessible to the general user, has a long road left to explore, and grows ever faster: it develops sooner than it is learned collectively. It is said that Newton developed differential calculus at the pace a high-school student learns it; AI can be faster. We are still in time to adopt it, since it remains in its early stage. The countries that lead it will dominate those that do not; the companies that embrace it will surpass those that do not.

That is where we set out toward interior and exterior design based on well-defined prompts — one of the most promising changes this technology brings to the sector.

The interaction between human and machine grows closer and may reach unimaginable limits, what is known as the singularity of AI, which will surely arrive by way of quantum computing. We could then cite Koestler and his Ghost in the Machine; but for now, let us make our architectural proposals more efficient with the tools presented to us in this disruption called generative artificial intelligence. It not only accelerates the design process, it also fosters innovation by enabling unconventional approaches. Architects and engineers can feed these systems with project requirements, constraints and aesthetic preferences; from that data, GAI can produce a variety of designs meeting the criteria of the firm or professional employing it — their style, their taste, and their way of thinking.

Conclusions

Generative Artificial Intelligence is unleashing a revolution in the architecture, engineering and construction industries, and of course in the economy. The collaboration between human creativity and the data-processing capacity of machines is producing more innovative, efficient and aesthetic buildings, and will reach the point of making them far more sustainable as well as profitable.

Being part of this disruption brings us closer to opportunities, but also to challenges about how to take advantage of this technology without losing the individual's artistic and intuitive capacity, nor ethics, nor one's own identity. But this synergy, within diffusion models, promises an exciting future for architecture, engineering and construction.