In today’s world, artificial intelligence has entered all aspects of human society and life. From manufacturing to medicine, self-driving cars to intelligent games many fields are being affected by this phenomenon. The biggest impact of artificial intelligence can be recognized especially in creative fields such as art, design (see fig.1), and architecture. In this essay, the impressions of AI will be explored specifically in the architectural realm.

Artificial intelligence can be defined as “The technology building smart machines able to perform tasks that generally need human intelligence in machines that are programmed to think and act like humans”. As Lev Manovich discusses in his book “AI Aesthetics (see fig.2)”, the dramatic increase in computer capacities, the ubiquity of digital devices and networks, and the challenges and opportunities brought by the ‘Big Data’ trend of the 2000s have also affected AI, we moved from automation of a single mind to a kind of ‘Super- cognition’ (Manovich, 2018, p. 8). 


Frieder Nake. Hommage à Paul Klee | The screenprint created from a plotter drawing produced by a computer program written by Nake. 1965

Manovich, L  (2019) AI Aesthetics. Strelka Press (figure.2)

Throughout ancient and medieval history, most of the architectural design and construction was carried out by artisans—such as stone masons and carpenters, rising to the role of master builder. All the tasks carried out by more than one ‘master builder’ and the buildings that are generated are a product of various personas. 

The beginning of “designer concept” started to emerge during the 20th century as the modernist style was being adopted more throughout different parts of the world. In this era, buildings that are generated were often attributed to and remembered by singular architectural figures. Therefore,  the authorship belonged to architects themselves (see fig.3). As the technological revolution struck, the field of architecture was subjected to change as well. The ease and accessibility that was brought by computational design tools started to replace conventional ways of design processes.


Outputs were created not only by humans but also these machine learning algorithms. Thus, intervention of AI in architecture appeared theoretically and practically. At the end, the design process and the product itself became quite intertwined between human and non-human species. In this stage, authorship of said creations should be redefined as a symbiosis between architects and AI. In her book Symbiotic Planet (see fig.4), Lynn Margulis (see fig.5) defines symbiosis as “simply the living together in physical contact of organisms of different species.” 

Symbiotic Planet (figure.4)

Lynn Margulis(figure.5)

Authorship in the conventional sense can be defined as “the state or fact of being the writer of a book, article, or document, or the creator of a work of art.” However, the arrival and generalization of AI processes in architecture urges to create a more inclusive way for the 21st century which is a symbiotic authorship. In that context, symbiotic authorship is the relationship between algorithms and human species in the Anthropocene era which could be for, against or neutral for both parties (see fig.6). 


Human-machine interactive design in architecture, particularly through the use of GAN’s in planimetries, should install a symbiotic mode of authorship. The full name of GANs is Generative Adversarial Neural Networks which is a type of neural network architecture for generative modeling using deep learning methods.

Stanislas Chaillou created ArchiGAN: a Generative Stack for Apartment Building Design that can use AI to analyze, generate and qualify floor plans. The network consists of two main pieces, the Generator and the Discriminator. The Generator transforms the input image to an output image and the Discriminator tries to guess if the image was produced by the generator or if it is the original image (see fig.7). Each step of the system corresponds to a Pix2Pix which uses a conditional generative adversarial network (CGAN) to learn a mapping from an input image to an output image to perform one of the 3 tasks – building footprint massing, program repartition and furniture layout (see fig.8) and lets the models learn topological features and space organization directly from floor plan images and control the type of information that the model learns by formatting images (see fig.9 an 10).





“In the case of floor plan layout, partitioning and furnishing the space by hand can be a challenging process as the footprint changes in dimension and shape (see video 1). The models prove to be quite ‘smart’ in their ability to adapt to changing constraints.” Stanislas Chaillou said it at the end of his essay. So we can see with the development of GANs deep learning can make AI smarter with remarkable flexibility in the foreseeable future (see video 2).

Viedo 1

Viedo 2

The intervention of machine intelligence has changed the original interpersonal relationship between architects and formed a decentralized network. Architects own the ability of information storage and design and architecture. In the post-humanistic context, a machine is not only a tool for construction, but also a tool for thinking. SO the machine learning algorithms in architecture introduce not only intelligence itself, but also the external extension of an architect’s ability and the architecture of a new signature system .




– Manovich, L  (2019) AI Aesthetics. Strelka Press.

– Tao (Nov 18, 2021) Lynn Margulis. Britannica.

-Chaillou (Jul 17, 2019) ArchiGAN: a Generative Stack for Apartment Building Design. Nvidia Developer.

 – Adjaye, Hirsch, and Otero-Pailos ( October, 2011 ) On architecture Authorship. Places Journal

 – Martínez Alonso (May 10, 2017) Suggestive Drawing Among Human and Artificial Intelligences. Nono.Ma

– Margulis, L. (October 8, 1999). Symbiotic Planet. Basic Books.

– Shrestha, A (November 13, 2019). Generating Modern Art using Generative Adversarial Network (GAN) on Spell. Towards data science.



Who is Designing?-Symbiotic Authorship in Architectural Planning is a project of IAAC, the Institute for Advanced Architecture of Catalonia, developed during the Master in Advanced Architecture (MAA01) 2021/22 by students: Aleyna Gültekin, Aaron Jude Pereira, Brian Woodtli, Fernanda Orviz, Mara Luisa Müller-de Ahna, Miran Calmanovici and Weihao Yin; Faculty: Manuel Gausa and Jordi Vivaldi