Digital technology, which once aimed to augment the realm of design, now seeks to overtake it. In the last year, artificial intelligence has been a topic of much discussion and debate. Workshops have been organised, and details of relevant mediums have been fretted over. A landmark moment for this rising anxiety was a conversation between ChatGPT and Neil Leach, British architect and theorist. The crux of the dialogue and Leach’s dire warning was that, according to the bot, architects might become a thing of the past. Truly, the world has changed with the proliferation of generative systems and digital technology claiming to revolutionise the realm of design by becoming designers. To deal with the complete ramifications of what this means for the future of practice, it bears thinking about a fundamental question, “What is meant by design”?

Design, or disegno, as defined by Alberti¹ constitutes the intellectual component of architecture, what the architect imagines and sketches, and is constructed by masons. This definition elevates the status of the architect to the artist. However, as Mario Carpo argues in his book The Second Digital Turn, the supposed golden age of digital design and technology blurs this divide, with designers who manipulate the created object via real-time simulations (what we commonly refer to as generative/parametric architecture)². As the architectural historian speculates, the rise of digital tools in design should have led to participatory modes of agency where the variable parameters for the design of an object would be determined by the designer, with their determination into form being controlled by an independent agent. Repeated several times, this model of parametricism indicated to Carpo layered authorship, blurring the idea of a primary author.

The humanistic tradition of architecture as art has often seen architects fight for control over the creation of form, and even in today’s digital age, they are hard-pressed to relinquish this privilege, as is evident by the presence of the starchitect figure. While parametricism seems to liberate the idea of the architect from a sole genius, the architectural expression of the idea is often (if not always) attributed to an individual architect and his intuition rather than technological development. This article is not an exhaustive history of digital in design, but the understanding of self-generation and its implications for authorship in architecture is important to discern the potential of an AI-generated reality. This is a future which now seems inevitable, with the question of autopoiesis of form emerging anew and the idea of authorship becoming muddied.

AI at its outset seems to take the idea of what Carpo calls open-source participatory design to its sci-fi conclusion, by giving people the power of design. With the combination of the right set of words, anyone can be a designer. If anyone were to create something akin to an architectural structure, the question becomes, what form will generate form? How will we determine what buildings we create if everyone can create them? There have been arguments for the adoption of AI in design that focus on how the use of words to generate form will lead to a more empathetic architecture. However, these disregard the mundane act of building by focusing on the experience of it. The act of building and construction, of course, cannot be divorced from architecture. Thus, the question of the AI future sits between the chasm of the most furtive debate of architecture: that of practice and theory.

Apart from the representative aspect of design, we must subsequently ask who will construct the dreams that androids dream. AI-generated design is currently rooted in the visual (such as the images produced by Dall-E, Midjourney and the like), where the thrill comes from how simple it is to create something resembling space at the click of a few keys. Should AI develop the power to do the work of an entire studio, making construction drawings, organising specifications and coordinating services, will the spaces it generates still arouse the same thrill? Or in their quest for perfection of structure, mechanics, sustainability (and so on), will they become the same boxy spaces we were trying to do away with? If we can generate spaces in a matter of hours, should we? If technology is the answer, what was the question, as Cedric Price opined at the dawn of the technological age?

This article has set forth a reflection on the means of production of architecture in the light of the rise of AI, by questioning what architecture itself means: is architecture to be just experience? Or is architecture something solid? That said, it bears thinking about what will happen to the modern professional once completely freed from the shackles of actually tracing lines that generate three-dimensional space. Like with the onset of CAD, will it result only in boring buildings and architects acting as managers? The question becomes more crucial when we understand that while everybody can create, not everybody will build. My longer thesis, part of an anthology on AI and the future of design to be published by CEPT traces the gradual devolution of the agency of the architect to speculate if the onset of the truly virtual means to topple the very ideas that modern architecture had based itself on the idea of the standard and that of the author. In the final essay, I critically examine generative systems of design and how they differ from or are similar to AI technology as a guide to understanding the future. A future that will be governed by how we choose to use AI today.

CEPT University’s annual essay prize focused on artificial intelligence and its implications for architecture in its last cycle (2022-23). The above presents an excerpt of one of the winning essays. The final volume will be unveiled on February 7, 2024, in a felicitation ceremony.

References

^1.The Renaissance theorist considered the author of the modern definition of the architect
^2.In The Second Digital Turn, Carpo shows how the “first turn” was characterised by mass customisation and the creation of free forms, the “second turn” was driven by optimisation and simulations. The “third turn” (AI-generated form) seems to be the Goliath to both these, generating free form as easily as writing this sentence.

(Disclaimer: The views and opinions expressed here are those of the author(s) and do not necessarily reflect the official position of STIR or its Editors.)