AI and technology Is generative AI (GAI) likely to impact the form of architecture? Mario Carpo contextualized GAI by describing how the modernist style developed from technological innovations in materials and their industrialized processes of fabrication. More recently, advances in computer-aided design and manufacturing have resulted in a stylistic pursuit of curvature and variation in architectural form. Mario argued that, historically, architecture has been influenced by technological change but not bound by it.
Mario Carpo: Sometimes, the history of architecture is changed by technology. And sometimes, the history of architecture changes without technology. If we look at Gothic architecture … European Gothic architecture was due to one technical innovation – not any new material, but a new way to deal with a material. Stereotomy [was] a new way to cut the geometric forms that made the building of Gothic cathedrals possible.
And so, if you lived in the Middle Ages, you could argue that the next change in architecture would come with a new technology. Wrong! Because when the Renaissance came, it was a major change in architectural style, but it was not related to any technical change. There was no new material, no new technology, no new anything that compelled Italian humanists to do away with Gothic architecture. They simply didn’t like it and they wanted to revive the architecture of classical antiquity – which, by the way, was much easier to build. So, humans went back to a way of building that was technologically way less sophisticated than what the Gothic builders were doing.
AI and style GAI’s unique approach to style imitation and transfer can reintroduce concepts that modernism seemingly abandoned. Mario noted that AI’s alignment with more classical ways of creating – derived from existing ideas, styles or works but adding something new – reintroduces imitation and style into the discourse in ways that are confronting and uncomfortable for architects.
Mario Carpo: For the time being, AI is purely a tool for the creative imitation of a dataset that we have created for the express purpose of imitation. And
