Can a Machine Perform the Work of an Architect? A Chat with Jesper Wallgren, Founder of Finch 3D
There has been a lot of talk about how automation will affect the way we do architecture, and what our role will be when technologies reach our own desks and work tables. In recent years, while we have seen how robotics and advanced technology are gaining ground in construction and manufacturing, new tools are emerging that promise to automate the design process itself. These would allow us to quickly and easily configure living spaces and their dimensions in the initial stages of a project, using simulations and artificial intelligence.
Will this automation be the future of architectural design? We talked with Jesper Wallgren, architect and founder of Finch 3D, to better understand this tool and its possible scope.
AD: Who is your target audience? Who could this algorithm help?
Finch is a tool for Architects that will automate repetitive tasks and guide architects through the design process by helping them making more informed decisions using simulations and AI. By doing this, architects will be able to spend more time creating great architecture and allow them to evaluate their design easier.
Finch works as an extension to already established CAD/BIM tools. To build it as an extension was a decision we took very early in the design of Finch. We think it’s important that architects can continue to design in the environment they’re use to – Finch will just make it smarter.
Finch is intended to be used starting in the early phases of a project. We usually say it operates from the first napkin sketches to the building permit. You might say it’s in the scale of 1:100.
AD: How/Why do you think this could improve architecture around the world?
I’ve been running my own architecture office for 5 years. In the early phases of a project, a lot of time is spent evaluating our designs, making sure they follow regulations, looking at zoning laws, and delivering the numbers the client need. This is very important since it’s in this phase that the most value can be gained but also lost.
Finch will give you instant feedback that makes sure your design is up for the task, and you’ll be able to focus on what you do best – creating great architecture.
It’s important to remember that Finch is only a tool, very much like a pen or a hammer. Finch will only improve architecture as much as the person using the tool is able to. But anyone who practices architecture knows that you’re often working towards a deadline. In many cases, you won’t have time to assess multiple options and to dig deep into every one of them. Finch will help you evaluate thousands of designs fast and provide you with the information you need to make an informed decision.
AD: How does the intelligence of your tool work and how has it been configured to deliver spaces of “habitable” dimensions? What kind of attributes and constraints can users add to these plans to better adapt them to specific projects?
Finch has two types of intelligence, rule-based and AI. Rule-based contains algorithms with inputs the user can change themselves. It’s applicable to things like building height, apartment distribution, and wall thickness. This is the type that will automate a lot of repetitive tasks. The AI part focuses on understanding our users and generating different design suggestions. The more you use the software, the smarter Finch will get.
There’s an infinite number of ways to design a good living space, but there’s also a lot of recurrent features. Most have a toilet, a kitchen, and some bedrooms. By analyzing existing dwellings, we’ve been able to learn a lot and generate suggestions based on that. As I mentioned before, the generated suggestions are not intended to be the final design but rather a great start for the architect to tweak and to add their own touch to.
AD: Does Finch consider factors like lighting, installation, or proper equipment for disabled users? (Considering these items could possibly differentiate automatic design from the work of an architect)
The user experience is a high priority for Finch. We’ve been focusing a lot on keeping Finch intuitive and fast.
With that in mind, Finch considers things like daylight and other elements that tend to influence architecture. Installation is considered on a more diagrammatic level than a very detailed one. For example, Finch might suggest that you need a shaft rather than telling you what kind of pipes it will contain.
AD: Some people have said that tools like this will replace the need for architects around the world. According to your vision, what will the role of architects be in the future?
The role of the architect should always be to create great architecture that takes advantage of the opportunities presented at a specific site or project. This is a very complex role where some decisions are based on facts, like budget and constructability, and others on softer values like what we consider beautiful and what are good ways of living. We believe that these decisions are best made by architects, and that Finch will help them make even more informed decisions.
As architects, it’s important that we take advantage of the new possibilities and tools that arise to maintain our relevance. This is a challenge that most industries face. AI and automation are getting better and can now preform operations we couldn’t even imagine 10 years ago. That’s amazing but can of course also be a bit scary. As an architect and a human, you might ask yourself, will I be able to perform as well as the machine? Sometimes the answer will be yes and sometimes no.
AD: What do you have in mind for the near future, in relation to your product and housing around the world?
Our focus the next year will be to release the first version of Finch. To do this, we’re continuing to work with our pilot clients to make sure Finch becomes a product architects love to use. Parallel to that, we will open a web-shop at finch3d.com where you can download a lot of the algorithms we’ve developed along the way.
The tools we use to create architecture influence the designs and the decisions we make. It’s always important to ask yourself if the tools you use are helping you create better architecture or if they are making it harder. After trying many different software programs over the last 10 years, my conclusion is that we need tools that understand us better, not the other way around.