How AI challenges architectural design - Archive ouverte HAL
Communication Dans Un Congrès Année : 2023

How AI challenges architectural design

Résumé

For several years, Artificial Intelligence (AI) has been spreading in all areas of activity. It aids in communication, movement, consumption, production, treatment, and training, and now, with a simple prompt, it suggests designs. This is made possible by an increase in the production of data and digital traces generated by our activities, which is augmented by the Internet of Things and what we refer to as big data. Being assisted by a machine, even if the term "Intelligence" is debatable, raises hopes and fears about ethics, values, and references. It has the potential to disrupt organizations and professional practices. As Computer Design professors in architectural schools, we must anticipate how AI will a!ect or act on the design process, as well as how students will use AI in their design explorations, whether they will hide it or develop a new workflow. We've seen how Pinterest and Instagram have influenced the search for graphical solutions, as well as how AI will challenge architectural design. Because it is not a matter of good or bad, we must look to the history of technology to see how computational principles emerge. We will look at Villars de Honnecourt's first measuring tools, Raymond Llull's work, Jacques de Vaucanson's automatons, and the birth of cybernetics to see how the concept of "perceptron" gives birth to artificial neural networks. We can see from these examples that issues of copyright, design ownership, and creation have already been raised. We will then create a short inventory of the "machine learning" techniques that have recently been implemented into tools that architects can use. We will also evaluate the most recent AI platforms, keeping in mind that we live in a fast-paced world with new breakthroughs in innovation almost every day. The calculated and generated representations open the door to new imaginaries, which must then be critically and consciously considered and apprehended. We will then refer to the first courses developed with various groups of students who, voluntarily or not, worked and used AI in the design process to see what AI can do for architectural design.
Fichier principal
Vignette du fichier
eaea16-2023-311.pdf (11.2 Mo) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte
Licence

Dates et versions

hal-04596700 , version 1 (16-07-2024)

Licence

Identifiants

Citer

Philippe Marin, Laurent Lescop. How AI challenges architectural design. EAEA16: Envisioning Architectural Scales in the Analogue and Virtual Representation of Architecture: Proceedings of the 16th EAEA conference, Royal Danish Academy - Institute of Architecture and Design - Copenhagen, Aug 2023, Copenhagen, Denmark. pp.249-258, ⟨10.60558/eaea16-2023-311⟩. ⟨hal-04596700⟩
182 Consultations
13 Téléchargements

Altmetric

Partager

More