Published in 2/2024 - Matter and Intelligence
AI Is a Tool for the Sustainability Leap
Artificial intelligence can benefit architects, especially when there are many variables in the design process – for example, reusable building parts from different sources, say Ron Aasholm and Maija Parviainen.
For the general public, artificial intelligence currently takes shape through text- and image-based tools, which also can be incorporated in architectural design processes. For architects, artificial intelligence has the potential, nevertheless, to enable a broader examination and renewal of the design process in building, when one builds the tools oneself and customizes them for the design process. At the same time, data can be used ever more extensively as the input for design as well as for analysing and evaluating design solutions.
Established commercial design and BIM software are all based on imitating analogue tools: we draw with virtual pens. Tools based on artificial intelligence do not have a corresponding analogue model, but rather work freely, with a new logic.
Increasingly more generative design software for architects is being introduced to the market. Often, these focus on creating multiple options of conventional designs, for example for new urban areas, by replicating the solutions of apartment buildings that are optimal for the housing developer. At best, they offer an opportunity to reduce the manual work associated with design. Similar components have existed, however, already for some time in the toolbox of parametric architecture, especially in optimization algorithms. Machine-learning algorithms that are more readily available than previously enable their further development.
One must consider whether the architect’s own design principles and practices can be turned into algorithms, and tried-and-tested design solutions and principles turned into data for the utilisation of artificial intelligence.
The important role of an architect in the development of AI-based design tools lies in asking the right questions and creating the criteria for software development. Without the inclusion of conceptual thinking in the design tools, the effects of artificial intelligence on the quality of architecture will probably be marginal, even though the productivity of an individual designer may increase thanks to the tools.
Artificial intelligence and the new tools will primarily affect the process, not the form or style. New tools should not be allowed to weaken design quality – on the contrary, tools should be built in such a way that they support the development of high-quality architecture. As a profession, we must not give this power away. When algorithmic tools are built, the design process, along with its assumptions, values and interpretations, must be broken down into parts – one must consider whether the architect’s own design principles and practices can be turned into algorithms, and tried-and-tested design solutions and principles turned into data for the utilisation of artificial intelligence. The development of AI-based tools can help us discuss exactly what assumptions and valuations the design process contains, and what is good enough in terms of quality.
Simultaneously with the boom in artificial intelligence, we have to find solutions to challenging sustainability issues: the planet’s carrying capacity cannot withstand continuous new construction and the extraction of virgin raw materials, irrespective of whether new housing areas are designed with artificial intelligence or not.
The generative optimisation tool matches existing building parts into building typologies defined by the architect and creates new volumetric options with a variety of combinations. Image: Studio MPRA
The construction industry’s transition to a circular economy, with its new data management needs, as well as calculators aiming for low carbon emissions create an interesting opportunity to integrate data into architectural design processes, so that we can increasingly better evaluate and compare different design solutions. We feel that planning tools are most useful in the early stages of the planning process of construction and demolition projects, where important decisions regarding, for instance, the building’s structural frame or the reuse of building parts, are made based on rough assumptions or insufficient information.
The use of algorithms makes sense in situations where there is a lot of data that changes during the process and which serves as the input data for the design – for example, reusable building parts from different sources. We have developed a generative design tool for utilizing reusable building parts, which in practice creates new plans for urban neighbourhoods based on the available structural parts and the design objectives defined by us. The tool produces different plan options with an emphasis on different features. If the building component data changes, then the plans can also be updated parametrically. The objective is to create possible future projections of what can be done with the existing building parts, and thus create value for them even before demolition.
For our software which is at the demo phase, we have given the exact framework for what we feel is possible in the given type of project, and we have determined what parameters we think we can use to create high-quality design. The demo therefore tirelessly produces alternative plans within the framework of our concept and quality targets. Although the main objective of the tool’s development has been to facilitate designing with reusable building parts, its byproduct was a software that creates options for urban neighbourhood plans based on building typologies and our own design library.
A similar program could tirelessly sift for parts suitable for new purposes from material flows or adapt new uses to existing building structures based on changing design objectives. The technological development of tools is important, but it is even more critical to consider what kind of software resources are used in such a development and what nails are hammered with these new hammers. If we are satisfied with automating the current design processes and feeding in the existing plans to be digested by the algorithm, then we will arrive at similar end results. but simply faster than previously. Artificial intelligence, however, can be a tool for the systemic change in the construction industry for the architect of disruption. ↙
RON AASHOLM and MAIJA PARVIAINEN work in the architectural studio MPRA which explores the relationship between algorithmic tools and design. MPRA’s project “Smart Design Application for repurposing used building elements”, has received support from the Low-carbon Built Environment Programme by the Ministry of Environment, funded by the EU.