In Conversation with Kent Larson (MIT Media Lab: City Science)
This transcript is an informal conversation between Sarah Fayad (MAUD ‘20), Shovan Shah (MAUD ‘20), Sai Joshi (MAUD ‘22) and Kent Larson on December 9th, 2019.
Kent Larson directs the City Science (formerly Changing Places) group at the MIT Media Lab. His research focuses on developing urban interventions that enable more entrepreneurial, livable, high-performance districts in cities. Larson and researchers from his group received the “10-Year Impact Award” from UbiComp 2014: a “test of time” award for work that, with the benefit of that hindsight, has had the greatest impact. The New York Times Review of Books selected his book, Louis I. Kahn: Unbuilt Masterworks (2000) as one of that year’s ten best books in architecture.
Sarah Fayad (SF):
Urban Design as a practice is changing with the integration of technology and also - how do you see the role of urban designers in particular and how it can be integrated with technology & science?
Kent Larson (KL):
The main challenge for urban design I think is innovation in modeling human dynamics. Urban design tends to be much more formal or dealing with infrastructure. But I think the big questions are, how do you build healthy communities? How can you increase the innovation potential? How can you improve health and wellness? Or safety and security? or How can you design a city that minimizes commuting by putting available jobs in sync with available housing that's affordable for work face, place and workforce and families, you know, those kind of more complex issues that relate to human behavior, I think, have to be the future.
It also needs to encompass public policy issues and incentives. The separate domains that exist in silos make urban planners think in a classic way to get a plan functional for the next 10-20 years.
How can we create incentives through tax laws or zoning, etc., for good things to happen? And then you have this whole world of technology, particularly in the Smart Cities world, where it's about optimization and efficiency of existing systems. But the reality is that in the future, I think there needs to be this harmonious integration of design public policy technology, so I think the number one hope I have is that the silos disappear. And technology will play a role in it. I think the most important use of technology, though, is not in the systems that are embedded in the city, like, you know, sensors and whatnot, but I think it's creating a more functional process.
SF:
Do you think, with the fact that the designers have this sort of multidisciplinary approach to problems, they will be major contributors?
KL:
I hope so. It's really the result of a lack of tools that allow people to do that, not a lack of motivation. And I think urban designers’ and architects are trained to solve problems, doesn't really matter what the problem is. And so, I'm hoping they will help lead the way.
SF:
In terms of technology, we have seen that it's very data driven, and you're getting a lot of data from cities, involving yourself and data from communications. Have you been limited in terms of data that's affected your research?
KL:
The data that is most important, again, relates to human behavior. So we're doing a lot with telecom data and mobile phone data, because that can give you a better understanding of the decisions that people make, how they move through the city, the places that they go and when and why. But it doesn't give you what motivates people. So for that you need to have the fusing of lots of different data sources. We use telecom data, we also use survey data like the national household travel survey. But, you know, they both have their limitations. We also need taxi data, bike share data, you know, all this stuff. And it's hard to get and even harder to fuse. So you can use survey data that gives you information about more of the values and the preferences, but an individual in that survey is not tied to individual behaviors in the telecom data. But for me, that's only the first step. In our city science work, we have a process that we try to go through that consists of five components. I don't call them steps because they don't happen linearly very often, but it's inside transformation prediction, consensus governance. And, what does that mean? How do things work today? That’s collecting and analyzing data.
You know what the current conditions are coming from institutions; in the evening where you want to get data from lots of different sources. And of course, you know, that indication can always have biases in it.. It's getting harder and harder. For example, the National household travel survey, the compliance 30 years ago was maybe in the 70%. You know, they mail out the surveys and they fill them up. Now it's dropped to, I forget what it was 15%, something like that. Tracking exactly political polling. So, it's very hard to get a representative sample when you're asking questions, you know about who you prefer in the next election right now?
So landlines have proved to be privileged over cell phones because it's hard to get people to answer questions on a cell phone and you know, it depends on your income and your education and then some people have political views they don't want to share with people. And a lot of urban science projects, starting in there today and on the big data, stuff that everybody's excited about. For me, that's just the very beginning because then you won't really want to go to the second phase of transformation.
Before you make decisions, you need to go to the third component, which is predictions. So, we have these ideas. An urban plan looks good. Some validated prediction models to try to predict the impact of those interventions. What we found some years ago was that even those three steps inside transformation prediction aren't good enough, you need to go to the fourth one which is consensus, because decisions are always made by a complex mix of stakeholders.
You know, people who live in the area, the politicians, the business leaders, etc. You want to bring them together in some process that leads to constructive conversations about future possibilities and then constructive action where people make real decisions about real problems. So, conversations are not enough. And the model I think that is used today, which is you hire a bunch of people in suits to make a presentation to the community, and then the people who were most opposed and angry show up and yell at them at an open mic, so that doesn't work anymore. We're trying to model and develop new processes for consensus.
And then the final one, governance, is just to say that design alone in the market alone won't lead to healthy communities. You need to have new systems in place to create incentives so developers will make decisions that are supportive of a healthy, high functioning community. And then you want to encourage the local people, the people to live and work there to behave in prosocial ways. We're looking at things like algorithmic zoning, with dynamic incentives that can be an alternative to this old, I think, obsolete process of zoning, which I think needs to disappear.
Let’s take Kendall square as an example. It is considered one of the best innovation districts in the USA. But as a community space, it is quite dysfunctional. There is almost no housing. Real estate is very expensive and only a few rich people can live here. There's no grocery store, pharmacy or community health care facility. You know, you'd have to walk a long way to buy broccoli if you live here for a toothbrush, and that is because real estate is so valuable that developers only want to develop corporate research labs or corporate offices or a little bit of luxury housing. So, cities have tremendous power over those decisions, potentially if they can harness the possibilities of creative land use regulations. And we already do that. So, there's the whole enclosed inclusionary zoning, which mandates that if you're building luxury housing, you must build a few units for poor people. So, if you just take that kind of static dumb process and made it algorithmic and you can then use all the stuff that people talk about, blockchain and smart contracts and public Ledger's etc. So, you could have dozens or hundreds of incentives to encourage property owners to make decisions that would then over time build out the infrastructure that's needed to create a viable community.
If you could take this you can say, okay, if you want to build research labs, this is the floor area ratio, that you're allowed to, that baseline that's existing zoning, you can stick with that. If you want. But we're going to allow an increase in floor area. Depending on what you choose to build out, you want to build housing for young professionals or families or the firemen and policemen that work in the district or whatever, then you'd have different incentives. And then you can execute that with a smart contract itself enforcing over time and flags go off if you violate that, so you don't have to have bureaucrats enforcing those. What I would like to see is to get rid of variances, because that's just a way for people that can hire fancy lawyers to game the system. The communities could meet and just twiddle these dials, with what and how much they need. The incentives then are fed into an algorithm. You start to build out enough family housing and you decrease those incentives and you can create incentives for workforce housing, or contributions to safe bike lanes or whatever. And there's no limit to what you could do.
There are lots of examples of incentives built into static land use regulations, but I haven't seen anything that's dynamic and algorithmic and responsive with real time data feeds. It could be automatic adjustments based on you know, the current conditions. So then if you do it right then as the economic conditions change, or new technologies are available, you know, suddenly, some new mobility system replaces Uber. Like the bike sharing app, you know that there can be a process of adjusting to the opportunities or the problems that are created with these changes that happen at a rate that are much faster than you can go through a whole rezoning process with which some takes years, sometimes decades.
SS:
Have you worked with private and public landowners on such kind of algorithmic based zoning policies?
KL:
I haven't talked to them in tempo. I've talked with people in Boston about this idea, you know, the Cambridge Redevelopment Authority. And I think in principle, everybody that I've talked to likes the idea, but the devil is in the details and we haven't worked out the details.
SS:
Going off that front the signals we want to understand how does the city science Media Lab impact policies and social issues in the public realm? And how does it vary from traditional practice? Is there any project that you believe has affected a city directly right now that has had the most impact?
KL:
Well, I think the one where we had the most impact was in Hamburg when we helped the city with the refugee crisis. And so, we started out with the idea that we work with the city on the 2024 Olympics, and then in the redevelopment of the Strasburg district in the port. And, they would have the influx of what they were expecting to be 80,000 refugees.
The mayor lost the referendum for the Olympics by two points. So then we shifted to the refugee crisis problems and the first thing we dealt with was the problem, but that wasn't how to build the housing. The Germans knew how to build the housing, the problem was working with the communities to identify the sites to build the housing. So, we went through a process where we used our cities to go platform with a local expediter facilitator and we did around 44 workshops with groups of 30 people and showed them how it would affect the neighborhoods if we moved certain units, with things such as the access to schools, jobs, mass transit, shopping, etc. And then feed in like it rather than arguing with me I could say, show me a better idea. But then, of course, the computational system didn't know all the local knowledge like this is in a park and there's a little playground there and you know, certain people use that. So, then combined with this real time feedback, plus a constructive conversation about all the other issues that no computational system can consider. It was released and I think we got pretty close. It was trying to use computational systems for what they're good at, which is calculations and then humans for what they're good at, which is, you know, exploring what the priorities and the values of the community would be.
Sai Joshi:
How are the processes here at MIT Media Lab? How do you trade them in a different setting, as in another country? How do you collaborate between technology and policy where it is just not available? How does one implement MIT Media Lab in another context without access to this groundwork of resources?
KL:
Well, come back in about a year because we're just starting some projects where we're going to be doing some work in informal settlements in Latin America where we're just about to start a project in Guadalajara and in Chile and Concepcion, Chile, hopefully. We haven't signed either of these agreements; we expect to sign them in December.
We're talking about a project in Nigeria, in a formal settlement where we're just putting together a project right now for the Venice Biennale with Norman Foster.
We've been working with him (Norman Foster) you know, now for about a year and a half and he's working on Odisha in India, and so we haven't gone through the community engagement process yet. But we have local partners that are involved in using traditional means. And we'll see how effective we should compete with some of these new ideas there. I saw an interesting challenge because I'm confident it can work in wealthy, educated cities in the north like Hamburg. You know, it will work in Guadalajara, we'll see. But I think well, you know, we have strong enough local partners that we'll figure something out.
Yeah, the intricacies in developing countries are easier said than done. Right? Well, what they're doing in Nigeria, will probably be working in the community with mapping. So, they're going out and if you look at the government maps, it's like a black hole there. But the key is that it is the job of the community and we can help them in that process. But yeah, they have to tap into their local knowledge that we're also looking at using drones to collect information from the air and then the local people can mark the specific points of interest in their community and their relationships and the flows.
SS:
Tell us about one dream project that you hope will be realized.
KL:
I’ve got a dream. There's a lot of them. I think I would love to take, you know, respond to this challenge in the three and a half billion people that are expected to live in informal settlements. I would love to see a process that would tap into all the new emerging population and abilities to transform those areas, you know, the air, their economies, the infrastructure, deal with the quality of life issues like sewage and water and power, but also look at new possibilities for local economies. And you know how they can become more a part of the larger economy and keep the wealth in the community. So, it's not like an architectural vision. I'd love to just see some process like that transform the lives of these people in those informal communities.