Can Engineers Build Inclusive Smart Cities?

Anthony Townsend
12 min readJan 12, 2015

Dr. Anthony Townsend

This speech was delivered on January 9, 2015 in Delft, the Netherlands, as part of Delft University of Technology’s 173rd Dies Natalis — full video available (begins at 1:25).

I want to talk to you today about a topic that’s often over looked when we talk about smart cities — inclusion.

Engineers are hard at work all across the world building these kinds of smart cities — from Amsterdam to Singapore to San Francisco. They’re rewiring and reprogramming our infrastructure for transport, water, and power.

But all too often, the benefits of these innovations are flowing to a privileged few. We now know that engineers can build smart cities. But can they help build smart cities that serve everyone?

Why does it even matter?

Income Inequality

It matters because inequality is the most important social and economic issue of our generation. Yet when this generation of engineers thinks about its public responsibility, what they usually think about is global warming. But just as sustainability has become everyone’s problem, inequality is quickly becoming so too. We should want to live in sustainable world, but it has to be a fair world too.

In the United States, we now have levels of inequality not seen since the stock market bubble of the 1920s. But the U.S. isn’t alone —it’s the whole developed world. Even the Netherlands, with its tradition of redistribution, wealth is increasingly concentrated.

This is a problem with no easy fix. It’s a result of the structure of the global economy we are living in. As the economist Thomas Piketty explains, the rate of return on capital, R, has been greater than economic growth, G, for some time now. This sets up a powerful feedback loop — since the rich have the capital, they accumulate wealth faster than the rest of us.

This is a relationship that will define our times.

Urbanization & Ubiquity

Now, it isn’t a coincidence that this wealth divide is happening alongside two of the biggest shifts in history — the rise of cities and the spread of computing. In SMART CITIES, I argued that 2008 was a watershed year for these trends — when humanity passed three critical thresholds.
That year, for the first time, more than half the world’s population lived in cities. We’re headed to 70 percent urban by mid-century, and 90 percent urban by 2100.

The second thing that happened in 2008 was that, for the first time, there were more mobile internet users than fixed ones. As computer scientists say, the internet is becoming ubiquitous — its everywhere.

The third thing that happened in 2008 was that — again for the first time — there were more things connected to the internet than people. About 5 billion of each. By 2020, the forecast calls for some 20 billion objects connected to the internet.

So we are living in an era of urbanization and computational ubiquity.

Urbanization and ubiquity.

The Network Society

These twin projects involve a massive expansion of networks of all kinds — roads and power lines, fiber optics and wireless — but also soft networks like social media and cloud computing. And it is the structures we are building in both the real world and the virtual world — a distinction that no longer really means much anyways — are what’s responsible for the way income and wealth are distributed.

None of this should surprise us.

It was all laid out very clearly in the 90s by the sociologist Manuel Castells who argued in a series of books that “the network society”, as he called it, was indeed an age of new connections — but connections carefully constructed or co-opted to reinforce the power of the institutions and groups building the networks.

Now, we’ve seen the power of new networks put to use by protestors and activists, but we’ve also seen it used by big corporations, government intelligence agencies, and criminal organizations, to extend their own reach and influence — perhaps far more effectively.

The Centralizing Internet

You can see this at work in the rapid centralization of internet infrastructure — which has been going on almost from the network’s earliest days — but has really accelerated in recent years.

This diagram was created in the early 1960s by Paul Baran at the RAND think tank in California. It’s the founding myth of the internet — a distributed communications network that would be far more resilient than the broadcast and switched telephone networks it would replace.

Many people think the internet still works this way. But for decades, the economics of scaling have actually pushed the other way. The internet is now massively centralized.

The cloud is taking this to new extremes. Take Google, which has swallowed a good bit of the internet into itself. Last year, something like 25 to 40 percent of all data traffic went through Google.

Exclusionary by Default

Now, this consolidation of power through networks is disconcerting enough. But as organizations like Google extend their networks into the physical world over the coming decades, the future gets scary really fast. Because what the internet of things is really about, is rationing resources — road space, electric power, clean water, etc.

The smart city will — to a great extent — will be exclusionary by default. Because as much as these new networks create connections — they introduce new rules to control those links.

This is the great downside of ubiquity. The smart city is full of barriers.
First you must be able to connect to the network. Then you must show your credentials by logging in. It’s like showing your passport to take a walk in the park.

Once you’re in, you need to know how to use it. Which is not always intuitive. Have you ever seen a senior citizen try to use a bank machine?
What’s important to understand is that smart cities exclude by design. It’s an inherent part of engineering these systems, their business models, and their regulation.

We see this in the rise of gated communities covered in blankets of digital sensors. We see it in new systems for predatory lending and predictive policing that are being used to exploit and suppress the poor and vulnerable.

Now, architecture and urban planning have always used physical structures to control the use of the city. But in a smart city you can simply write code. Smart infrastructure, big data and powerful new algorithms let us create new chokepoints and new enclaves far more effectively.

The scariest part of this future is that the gates of the smart city will be invisible. We may not ever know how or why or by whom we are being controlled.

What Can Engineers Do?

So, many of you are probably scratching your head, thinking — isn’t this a matter of policy? Why should engineers be concerned about exclusion in the smart city? What can they even do?

The truth is that engineers have more influence over our visions of the future than they have in a long time. Just look at the visions that are being put forward by technology companies about smart cities — they will be measured by sensors, analyzed by predictive algorithms, and remotely controlled by the actuators they trigger to reshape the physical world… if they get what they want it will be nothing short of an historic transformation of the material basis of our civilization.

But where’s the rest of the story? That’s what I kept asking myself as I was writing my book. The visions of the future of cities that engineers have tried to sell have focused too much on efficiency, security, and traffic congestion. They have ignored the social and ethical dilemmas I’m talking about. They were more concerned with controlling people than freeing them.

Efficiency and sustainability are important. But we can’t sacrifice democracy and equality in their pursuit.

It wasn’t just the goals of these visions that bothered me. The tools for getting us there too didn’t make sense.

In urban planning, there is a well-established notion of “wicked problems” — they aren’t unsolvable because of technical limits or a lack of data. They aren’t solvable because they are so big, and complex, and wracked with political conflicts that stakeholders can’t even agree what success looks like or how to measure it.

Yet the smart city promised that given enough data and enough processing power, we could simply compute solutions to any problem. This struck me as, at best, lacking in an understanding of how cities work. At worst, it was a disaster in the making.

See: Joe Flood, “The Fires: How a Computer Formula, Big Ideas, and the Best of Intentions Burned Down New York City—and Determined the Future of Cities”, Riverhead Trade, 2011.

Fifty years ago, in the 1960s, the first attempts to use computers to predict urban policy were undertaken. The results were almost always useless, and in a some cases very harmful. In New York, the city closed a number of fire stations based on the results of a deeply flawed simulation. The result was a wave of unchecked blazes that displaced some quarter million people from their homes.

We have more data and faster computers today. But does that simply mean we can make bigger and faster mistakes?

So, while I took issue with the way engineers were framing the problems of cities, and asking some hard questions about the tools they proposed — I also saw the tremendous assets that engineers bring to the urban challenge. And so my work has increasingly come to focus on understanding how to navigate that balance.

The context for smart cities has changed for the better. 2008, the birth of the smart cities movement, is nearly a decade behind us. City leaders around the world have a much more informed understanding of what smart technologies can do.

But they have hard choices to make. They have limited funds. And they need to make sure the benefits are widely shared. We have to see the city in a much larger picture than just efficient infrastructure.

So where do we start? How can engineers begin to incorporate inclusive thinking into their work? How do they make this a part of their responsibility to society not just to deliver new technologies and designs, but to protect the public interest as they do so.

Toolkits for Inclusion

Well, you aren’t starting from scratch. People have been thinking about this for quite a while. In the 1990s a global movement of researchers and activists that calls itself ICT4D — or information and communications technology for development — arose. They started by designing and exporting entire technology solutions for poor communities. They have slowly evolved to focus more on co-creation. And they aspire to actually empower the poor to build their own solutions.

A Planet of Civic Laboratories: The Future of Cities, Information, and Inclusion

And so there are a growing number of toolkits for thinking about how you design more inclusive smart cities. One I developed in 2010, which you see here, was a ten-year forecast on cities, big data, and inclusion. It described a dozen solution templates — things like anticipatory health care, local social commerce, on-demand resilience and others — that were opportunities lying at the intersection of technological and emerging social and economic trends.

With support from the Rockefeller Foundation, we’ve distributed almost 5,000 copies of this map and its been used to frame strategic thinking about national smart city policy in the UK, Mexico, and most recently India.

Three Areas of Focus:

Open information architectures, Transparent and interpretable models, Engineering ethics education

The title for this speech is — Can engineers build inclusive smart cities?

And the answer for me, is yes. But to do so, we need to embrace approaches that empower communities and build capacity for innovation. We cant simply deliver shrink-wrapped solutions based on our own assumptions.
So with this in mind, three areas jumped out where I think engineers can advance the smart cities movement in a more inclusive way.

The first area is the infrastructure for urban data and computation. We need to confront the exclusionary nature of smart cities head on. We need to make data architectures open to drive innovation. Modularity and inter-operability are key — so that anyone can build new applications on top of at little or not cost.

But all too often we build systems that are proprietary, too expensive, and not extensible or scalable enough.

Often this is done in the name of security. But closed isn’t a guarantee of safety, as we’ve seen in the last year. We can do both open and secure.

The digital architecture of our smart cities needs to be less like a mainframe and more like the web.

The second area is about how we integrate automation and analytics into urban government. Here, we need to stop building black boxes — and instead build models that are transparent and interpretable.

By transparent, I mean that the assumptions embedded in simulations and predictive models, automated infrastructure, and decision-support systems are open to public scrutiny.

All too often, this code is hidden from view.

By interpretable, I mean that machine learning techniques have advanced to such a state that the way they produce results is beyond human understanding. People don’t trust black boxes — they prefer models that emulate the way human experts think. We are seeing tremendous apprehension among doctors as predictive models are introduced in medical diagnostics — this is a sign of things to come when they are used to make public policy.

The final area is the role of ethics in engineering education. Ethics is only just taking hold in engineering and business schools — in the United States, the latest accreditation guidelines for engineering schools — for the first time — requires students receive merely “an understanding of professional and ethical responsibility”.

But the smart city is an ethical minefield, and it will only get worse. Sensors that collect private personal data to manage the city for public good. Ownership of data and its re-use by third parties. The unintended consequences of automation. And of course, the trade-offs that must be made to minimize exclusion.

I wanted to talk about inclusion in the smart city, because I believe it is a huge missing piece of the smart city movement.

But the Netherlands and Delft are in a unique position to fill this gap. You can build on the great Dutch legacy of social and economic equality. You can build on the great Dutch traditions of city and regional planning. You can build on the great Dutch tradition of trade to create new ways of citybuilding and city-retrofitting that can be exported to the rest of the world.

Delft University of Technology

And this is where I would hope to find a way to work together in the future. Engineers need to do more than design things — they need to get involved in how their designs are legislated, how they impact the planet and their communities.

While smart cities might not be able to fix the structural inequality that Piketty highlights. But they could certainly make it much worse. It’s part of our job to prevent that.

There are two areas of research, that I’m embarking on over the coming years, that I think can connect to what you are doing here.

The first is practical — how should cities plan for technology, and how should that be integrated with all the other kinds of long-range planning they do already? Dozens of cities around the world have published, or are developing, technology plans. If there is an emerging practice of “digital master planning”, how should it be done, and by whom?

The second topic is more speculative — what should a new science of cities look like? The newly founded Amsterdam Institute for Advanced Metropolitan Solutions is one of a number of new urban research centers that seeks to study cities through the lens of big data.

But where is this new movement headed? What new knowledge will it produce, and how will it be applied? How will it connect to existing urban research communities? On a fundamental level, what does it mean that physicists, statisticians and computer scientists and others are going to have a big hand in shaping 21st century urbanization? How will the planners, the architects and mayors put them to good use?

The big picture is that over the next 50 years, humanity will build for itself to a system of cities we’ll need to live in for centuries to come. We need to get the recipe for inclusive city-building right before too much time passes.

And that means involving everyone in its creation. And I can’t think of anyone who expressed this more clearly than the great urbanist Jane Jacobs, when she wrote:

“Cities have the capability of providing something for everybody only because and only when they are created by everybody.”

Thank you.

This speech was delivered on January 9, 2015 in Delft, the Netherlands, as part of Delft University of Technology’s 173rd Dies Natalis — full video available (begins at 1:25).