AI in urban planning helps improve design and outcomes

Using AI in urban planning can help designers and city planners find the best design solutions that will benefit everyone.

Autodesk Video

May 18, 2022

5:33 min. video

View transcript

Brian Pene, Distinguished Research Scientist, Autodesk Research Programs: Urban planning is really the ability to create urban spaces, and it’s not just incorporating one or a few things; it’s a really complex system. So it’s incorporating things like the environment, weather, buildings, systems, economics, supply chain, traffic—all of that different information. How can you design a city so that it’s not only livable better but it’s better for the environment, it’s better for the planet?

Ray Wang, Architect and Principal Research Scientist, Autodesk: In urban-planning projects, you get a lot of data; you have a lot of community involvement. And a lot of times, it’s really tricky to balance those different goals and objectives. So what AI allows us to do is to help find a common place to put that data and have that data be available to all these stakeholders and also use that data to create new types of designs and solutions that a human by themselves might not be able to come up with.

Carl Christensen, Vice President, Unified Design, Autodesk: The reason the potential for AI is so huge in architecture is that there are so many hard and complicated calculations, simulations, analyses, and consequences that are interconnected and intertwined and limit the stakeholders from expressing and utilizing their creative potential, especially with the urban growth we’re having today with creating better cities that are kind of livable and more sustainable but also more efficient in terms of actually housing the people that are moving to the city. It’s just so complex.

Håvard Haukeland, Co-Founder of Spacemaker and Senior Director, Autodesk: The way this works today is that humans come up with a design proposal and then check how the design performs. We’re turning that upside down so that you can be explicit about what kind of challenges you have on the site and what kind of outcome you want. And then you can get support from the computer and the software in the system to help you find out how to get to that solution in the best possible way. It’s kind of an assistant sitting on your shoulder trying to help you get there faster—but also to discover the best possible solution on the site.

Pene: The one thing that humans do well is framing and reframing problems and being creative, and the thing that AI does really well is the really manual, laborious stuff of looking at thousands and thousands of different permutations and parameters to come up with optimal solutions.

Christensen: For humans to handle very hard problems that have a lot of consequences, they need to simplify it. What we generally see is that we unconsciously choose some paths within the kind of realm of the possible. You need to start somewhere, so you start using pure intuition. But that pure intuition is, of course, based on biases, and those quickly get built into the foundations of a development project.

Pene: Humans have bias. I mean, it’s baked into us, so engineers have their own biases. Designers, architects have their own biases. So talk about AI, how can it help? It’s taking away the individual bias and incorporating a collective bias to serve the greater goals of all. It’s really about looking at outcomes by incorporating lots of different inputs. You know, like in the cities where they have lower income housing, those areas have less trees. Just by a sheer fact of less shade, you get sweltering heat at certain times of the summer. Every single part of a complex system has the ability to have things happen based on small parts of that.

Wang: When AI is facilitating, let’s say, a discussion around an urban-planning project, it makes sure that every piece of that is captured, that the community voices could have as much weight as developer voices, as ecologist voices, for example. So it allows us to balance and meet everyone’s goals.

Haukeland: People often ask us, if we introduce AI into this discipline, wouldn’t design then be more standardized? But it’s actually the other way around. Computers are able to think about more things than humans. They can do calculations much faster than us. And if you combine the human way of thinking with computers doing the math faster, then we’re actually able to come up with more complex solutions and complex outcomes.

Pene: Diversity creates innovation. We can make humans more aware of outcomes about a small and consequential decision I made here that all of a sudden there’s a bigger outcome over here, then, oops, you know, I should have thought about that.

A lot of this is just about creating awareness around those things. And that’s the things that AI can help by simulating things that we probably don’t have the capacity to do or looking at options in a way and then bubbling up an insight to a human that uncovers a blind spot.

Christensen: I think we’ll see a huge volume and a huge change in how people work with these projects, being able to focus more on outcomes and letting the machines do more of the heavy lifting, more coming together around the important questions of how to create better cities.

Haukeland: So what excites me the most is that with this technology advancement that we’re just in the middle of right now, it will be more fun being an architect again, and we can actually have a significant impact on the built environment.

Pene: What if you could do a lot less of the manual, laborious stuff, and you could focus on being more creative? What happens when you can do that and think differently about your design, and it still benefits your goals? But maybe it even benefits the goals of others.

Christensen: You get better cities and better projects, and those projects will persist in the built environment for decades and possibly hundreds of years. So I think the manifestation of AI in cities is that cities will be more human and more optimized for humans in the years to come.

Recommended for you