Local intelligence: time to learn from AI

Abstract

AI research in architecture is flourishing, and there are plausible and praiseworthy experiments using generative models. These experiments could result in intelligence that uses architectural knowledge and opens new learning opportunities, although guidance is still required in this area. With the Local Intelligence (LI) framework, we hypothesize a web of distributed networks to connect different forms of knowledge linked with architectural context. We assume that the tacit knowledge of vernacular architecture corresponds to the implicit knowledge of folklore music when the anonymous designer and the user are the same- the local people. With a multimodal AI model, we call 'music2architecture' - 'architecture2music', we argue that sharing the same locality/localness may lead to the emergence of a typical, previously hidden pattern (of wisdom) that we can learn from.