Abstract
The notion of agency in architecture and urban design is increasingly characterised by non-linearity and complexity. As Artificial Intelligence (AI) models become more sophisticated, authorship and control over
the design process are increasingly difficult to trace. The agency that underpins the design and making urban design and architecture is today characterised by a number of qualities, including trust in algorithms and automation, growing levels of complexity, contingent elements and indeterminacy.
Within this idea as a main framework, we explore possibilities of designer-machine co-design, specifically focusing on conditions of emergency and disruption. To do this, we present two related projects
developed by our team in the past 2 years. In the first one we developed a quantitative method to find
successful patterns in communities’ daily routines to assess and improve their resilience based on
physical urban elements using position, density and mutual proximity of urban typologies (Carta et al.
2021).
Following on the result of the first project, in the second we used AI models (object detection and
computer vision) to assess the extent to which communities would respond to change and disruption
(slow-pace as in climate change, as well as fast-pace as natural disasters). Our model predicts the
resilience of communities on the basis of urban morphology and configuration, using satellite imagery
(Carta et al. 2022).
With these 2 projects, we would like to show and comment ways in which computational design methods
are used by architects and planners to observe, analyse and measure the resilience of urban
communities, as well as to predict and suggest ways in which it can be improved.
In the analysis of these two projects, we focus on the interaction designer-machine, building the
argument that human and algorithmic intelligence should be considered as a whole, where the parts
complement each other on the basis of their individual strengths and idiosyncrasies
the design process are increasingly difficult to trace. The agency that underpins the design and making urban design and architecture is today characterised by a number of qualities, including trust in algorithms and automation, growing levels of complexity, contingent elements and indeterminacy.
Within this idea as a main framework, we explore possibilities of designer-machine co-design, specifically focusing on conditions of emergency and disruption. To do this, we present two related projects
developed by our team in the past 2 years. In the first one we developed a quantitative method to find
successful patterns in communities’ daily routines to assess and improve their resilience based on
physical urban elements using position, density and mutual proximity of urban typologies (Carta et al.
2021).
Following on the result of the first project, in the second we used AI models (object detection and
computer vision) to assess the extent to which communities would respond to change and disruption
(slow-pace as in climate change, as well as fast-pace as natural disasters). Our model predicts the
resilience of communities on the basis of urban morphology and configuration, using satellite imagery
(Carta et al. 2022).
With these 2 projects, we would like to show and comment ways in which computational design methods
are used by architects and planners to observe, analyse and measure the resilience of urban
communities, as well as to predict and suggest ways in which it can be improved.
In the analysis of these two projects, we focus on the interaction designer-machine, building the
argument that human and algorithmic intelligence should be considered as a whole, where the parts
complement each other on the basis of their individual strengths and idiosyncrasies
| Original language | English |
|---|---|
| Title of host publication | RGS-IBG Annual International Conference 2022 - Proceedings |
| Pages | 466 |
| Number of pages | 1 |
| Publication status | Published - 31 Aug 2022 |