The computer has already fundamentally changed the role of the architect once before. CAD design software, described as “the greatest advance in construction history,” hauled architecture into the digital age in the 1980s. Artificial Intelligence (AI) promises a second revolution. But what will be its impact? If we break down the role of an architect today, it raises questions over which steps have potential for future automation. Could project-related research be produced by a machine? Could a digital simulation accurately describe a project’s economic opportunity and replace the role of a financial modeling team?
Architects since Plato have based their designs on mathematical principles, and data has remained at the heart of Architecture teaching since. The Neufert, a reference guide for spatial requirements, has arguably been the most influential book in shaping architectural design since its release in 1936. More recently, parametric design, whereby a computer generates a range of design options based on a set of parameters, has been widely adopted to help generate design options. For engineers, algorithmic automation has been at the heart of industry conversation since the 90’s. Engineers use algorithms to minimize the impact of heat gain on a building, to define the materials they use and the shape of the cantilevers they need. Simulations can describe the best road layout based on tracking data in cities, where the flow of people and vehicles can be measured and used to inform optimal layouts.
The use of AI promises to build on these foundations and provide a step change in the way we work, improving fundamental design quality whilst enabling us to design in a way that’s more responsive to the changing needs of a building. Conceptual research has gathered pace and is beginning to be interpreted on a pragmatic level. Last year, the Harvard School of Design demonstrated how AI could successfully generate floor plans and suggest plausible design options based on its findings. Researchers at the Why Factory have developed a Block Maker tool that uses algorithms to explore how AI can power modulated buildings which respond to the users’ changing needs.
Another example is SpaceForm, a data driven design tool currently being developed by my firm, Squint/Opera, in collaboration with UN Studio and BIG. This tool allows architects and planners across the world to collaborate in real-time in virtual reality and use data to influence decision making. The user can interact with scale models on a virtual work table or immerse themselves at 1:1 scale – allowing the entire project team to experience the consequences of their design decisions. Data provided by the platform will provide a range of feasible design options then help designers adjust and optimize design decisions accordingly.
While the pragmatic and technical challenges of architects are being addressed, the question remains whether AI can reach the creative requirement necessary to achieve high-quality design. There are encouraging signs from adjacent industries that computers could possibly generate creative solutions in the future. Paintings by Google DeepDream are becoming increasingly convincing, and in 2017 a classical composition by AI left the audience unable to differentiate between human and machine composition.
So, what approach can an architect take towards this new technological development? One option is to ignore it and leave the task of adapting to technology to the next generation. Option two is to incorporate technology into the design process. Architects can harness new software platforms, like SpaceForm, to make their jobs easier across various parts of their role, from optimizing designs to improving their workflow.
Option three involves this substantial change in thinking about what it means to be an architect. Currently, the role of the architect is to design and deliver a project. But the data that buildings create can and should be used to continually enhance how a building optimizes its function. Information, like when rooms are actually being used, where most people are getting sick, or where the building loses the most heat, can turn into insights. Data can be analyzed and harnessed in order to provide new design solutions for architects, allowing them to make improvements to the design throughout the lifecycle of the building. This approach will change how we think about designing buildings – with adaptability and change in mind.
As a trained landscape architect, flexible design is at the core of my field. The purpose of public space is less defined and is used more extensively and creatively. You can drink a beer in the playground, create your own route across the park, exercise in a park or work there. The only constant in landscape architecture is changing, of the seasons of the users and of the space itself—at the core good designs are made to be adaptable over time.
In contrast, buildings are currently designed with an ‘end state’ in mind. Architects create new structures which decay over time. Looking at the life cycle of a building, our new challenge is to think of design as a continuous process with no end. We should design buildings like we do software: agile, on-going development and constantly optimized based on user data and analytics. This will allow buildings to become more resourceful, efficient and flexible, which most importantly can provide a tailored experience catering for the evolving needs of its end-users.