As the Chorus of Dumb City Advocates Increases, How Do We Define the Truly Smart City?

By Stephen Goldsmith • September 16, 2021

Recently, the Boston Globe published a provocative interview with Shannon Mattern, who asserted that smart cities are, in fact, dumb. Mattern, a professor of anthropology at The New School for Social Research in New York, joins the calls of several others who are advocating that we throttle the smart city enthusiasm.

In a more nuanced opinion piece, Shoshanna Saxe, an assistant professor of civil and mineral engineering at the University of Toronto, advocated in the New York Times that often, ‘dumb’ cities will do better than smart ones, setting up a false choice between people and technology. According to Saxe, “the parks, public spaces, neighborhood communities, education opportunities — are made and populated by people, not technology. Tech has a place in cities, but that place is not everywhere.”

Clearly, in writing this response, this author has a specific point of view—this response, after all, is posted on a site called Data-Smart City Solutions, and this author manages a program for city chief data officers. While the critics of smart cities raise legitimate questions and concerns, they also raise certain strawmen – the failed Sidewalk Labs effort in Toronto being their highest profile.

First, we need to deal with the definitional problem—there is no clear definition of a smart city. Years ago, as an elected prosecutor, my team in the district attorney’s office implemented what were then the country’s most advanced digital tools to increase the child support we collected from noncustodial parents and distributed to struggling mothers. Collections surged from $900,000 to thirty-eight million dollars annually; to me that was a harbinger of smart city efforts. When, as mayor of Indianapolis, we used digital geographic information systems to identify the city’s most neglected neighborhoods, along with their assets and needs, that struck me as a smart idea. Especially since these analytic visualizations sparked and provided the foundation for community partnerships, a substantial investment in services, and hundreds of millions of dollars in infrastructure.

And, when I worked as Mike Bloomberg’s deputy mayor in New York City, we took the first steps to create a data analytics center that would infuse data-driven decision making into a wide array of areas, from inspections to tax collections, ensuring safer restaurants and increasing tax revenue. Bloomberg’s Deputy Mayor of Social Services Linda Gibbs introduced one of the country’s most advanced uses of digital tools, which simplified and integrated access to social services for those in need. To me, all of these are examples of a smart city; technology was implemented at the local government level to improve the lives and experiences of residents.  

To move away from the strawmen—the arguments that a city can’t pay attention concurrently to people, equity, and technology, or the one that holds up Sidewalk Labs’ ambitious and now largely abandoned effort to develop a smart city enclave in Toronto (which was much more a commercial undertaking than a city undertaking), we need a definition. While no simple task, I propose that a city administration could consider itself smart to the extent it adopts these ten criteria.

1. Uses data to deliver city services based on where and when they can do the most good.

Most cities still operate on routines. For example, in these routines a call center receives a request for service; an operator logs it into a system, which eventually leads to crew being dispatched. But what if the traditional routine reinforced unequitable service delivery and dispatch? In Baltimore, an analysis of 311-reported cleaning work identified that this process resulted in geographical inequities in the way sanitation crews were being deployed. With this data, the Department of Public Works implemented a new sanitation strategy that focused on “clusters,” which allowed the crews to be more efficient and insure that overlooked communities received better services.  A smart city uses data to determine priority, identify systemic problems, and understand what caused the issue. In every area of government there are outliers and “frequent flyers,” and a truly smart city will determine the causes of those frequent concerns and try to solve the underlying problem.

2. Creates digital platforms that allow it to use IoT data to improve the way it builds, maintains and uses physical infrastructure.

A smart city builds digital infrastructure that enables the integration of Internet of Things (IoT) data to deliver better, and safer, cities. It uses these digital platforms to better construct, monitor, and maintain infrastructure. Thanks to smart sensor technology, a bridge in need of attention sends a signal long before it collapses; a neighborhood with dangerous particulates in the air sends an alarm to public health officials. Mobility officials can see – in real time – curb usage by type: delivery vehicle, car idling, or scooter parking, and adjust the allowed usage and pricing to address demand and congestion. A smart city is not defined by the amount of its IoT data but rather its use; how it analyzes and uses the data to improve the quality of life and safety. Further examples of smart infrastructure technology can be found in Building Back Better with Intelligent Infrastructure.

3. Makes public employees smarter in their work.

In too many cities public employees are stymied by narrow job descriptions and outdated tools. In a smart city, all these dedicated public employees will be given the training, tools and technology necessary to improve their work, and therefore improve the city at large. Smart cities can also use machine learning to resolve simple requests, thereby freeing employees so they can address more complex problems. For example, machine learning can handle a large number of 311 calls, increasing the responsiveness of operators to calls that need more research.

4. Enhances the way it listens to and involves the public.

A smart city involves the community in deeper and more comprehensive ways than before. It will expand the definition of community meeting to include digital meetings and virtual design charettes. It will use anonymous sentiment mining to understand what specific communities think about an important issue. Recently, we participated with the Robert Wood Johnson Foundation on a project around vaccine hesitancy.  City officials wanted to better understand the concerns that certain communities had around COVID-19 vaccines. Using anonymized social media data, city officials in the 18 cities across the U.S. were able to understand and address the hesitancy that many residents had, using the concerns that the residents themselves publicly shared and discussed on online platforms. The social media data revealed issues that the mayors could immediately address, including difficulties in accessing vaccine registration appointments in communities of color and confusion over side effects.

5. Uses digital tools to operate more equitably.

A smart city uses data and visualization tools to understand the delivery of services—quantity, quality, and responsiveness – by metrics like poverty and ethnicity, in order to identify and correct for systemic inequities. Food deserts or urban heat islands will be identified through data visualization tools, often uncovering historical, intentional discrimination, and help shape planned remedies. Transit data will be used to improve access to work and needed services, particularly for women and non-binary riders. Crucially, smart cities will have folks dedicated to not only monitoring possible biases in algorithms, but also how to use data to identify biases in the daily actions of public employees.

6. Addresses sustainability and resiliency.

Digital infrastructure and tools facilitate dramatic breakthroughs in sustainability. A city can reduce its CO2 and particulate emissions with IoT enabled advances that reduce idling at turns and queuing at intersections. Analytics and IoT data can organize commercial loading zones that make it easier for delivery trucks to park rather than burn fuel waiting and telegraph open parking so that motorists don’t cruise looking for a spot. Smart cities are not only able to identify and eliminate dangerous emissions that contribute to climate change, but they’re also able to handle disasters. As cities across the globe experience more extreme weather, sensors can help cities understand and preemptively address flooding and combined sewer overflows, not only saving lives during a crisis but also preventing future contamination of area water supplies.

7. Protects privacy, security and transparency.

A truly smart city has explicit privacy and security policies that apply both to its own operations and to those of its vendors. It prioritizes anonymity and ensures that all policies, monitoring, and data collection are transparently shared. The privacy rules applicable to data collected on public spaces differ from those on private property but the distinctions between the two lacked clarity in Toronto.  A truly smart city takes seriously its responsibilities to secure and protect the privacy of its residents. A smart city has the same duties to protect and care for residents as any other city; there have (and continue to be) very real lapses in those duties, but this fault does not lie only at the feet of technology. 

8. Acts in real time.

A smart city uses data, including IoT data, to help it act in real time, immediately identifying and working to resolve a problem as it arises. During the COVID-19 pandemic, cities were faced with outbreaks that were difficult to trace and contain, particularly as infected residents were not always symptomatic. Municipal governments across the world turned to wastewater monitoring, as the COVID-19 virus could be detected in sewage several days to a week before an outbreak was reported. Data accelerates the responsiveness of a smart city, expediting mitigation and prevention.

9. Focuses on its residents.

Smart cities put residents in the center. While one hopes and believes that this is the aim of all cities, the technology of a smart city facilitates a resident-centric orientation to a degree previously unimaginable. Smart cities create digital omni channel customer service. They use smart systems to collapse time-consuming bureaucracy. They reform how a customer applies for a permit or a license. Digital tools allow the smart city to convert sequential steps to concurrent ones, saving time and money for the resident and the city administration. Looking to get a construction permit for your home? Now there’s no need to visit multiple agencies; residents can easily apply for a permit online. In Miami, applicants  can access over 200 services directly through the city’s website. A smart city puts the user experience at the center, in a way that is faster and easier than ever before.

10. Adopts a new culture and organization.

A smart city invests in the right staff, not just the right technology. The Lab @ DC, the District’s evidence-based innovation team, partners with the city’s operating agencies to improve benefits delivery. Internal training academies can teach interested employees how to use data to improve their work.  A smart city celebrates, and supports, employees who used data to unlock a new way to solve a challenging issue.

With those parameters guiding our definition, let's return to the rationale of the “dumb city” advocates, starting with a note about Sidewalk Labs. This project presented a unique set of questions in Toronto, which brought to light both benefits and risks of this technologically-based development. The original project involved the real estate development of several acres along the waterfront. The now-defunct project, called Quayside, introduced an important set of questions about the role of the city in setting rules and demanding transparency when it authorizes large scale building projects, especially when the line between public and private areas is blurred. For now, however we will be setting that one example off to the side as less than germane to the central issue of smart city government and a rather unique construct in the smart city debate.

Our rebuttal to the critics centers not so much on their general concerns, which raise legitimate questions, but rather that they present false choices. For example, the Globe article argues that “marginalized individuals are often either insufficiently integrated into smart systems, which means that they don’t benefit from its efficiencies, or are overly represented within its surveillance and security apparatus.” These complaints are not about technology but rather policy. To what extent a city will use cameras and face recognition is a policy decision that brings with it difficult tradeoffs. Algorithms can compound bias, but that bias originates in the data of the dumb city on which the algorithm trains. Staying dumb allows the continuation of these day-to-day discriminatory practices. A smart city monitors algorithmic development while concurrently utilizing data to uncover unfair practices.

Mattern also argues that “the smart urbanism assumes a technical solution for problems that are often entangled in socioeconomic and cultural factors.” As discussed in our definitions, a smart city centers itself on its residents and utilizes tech tools to expose unfairness and mitigate the challenges faced by neighborhoods or individuals systemically ignored. A dumb city would be less able to solve these problems without the clear picture of circumstances faced by individuals in those communities and their real-time voices that express the frustrations and experiences of those living there. If the smart city critics actually want to assist these residents, will they really be better off with existing transit services that impose a serious cost on poorer residents in terms of time and price, or would a smart city redesign include subsidized access to other first and last mile providers, like pooled Ubers or Via vans (as in South Bend) or using data to guide the distribution of subsidized transit passes (as in Boston)? 

Another category of complaints presented by Mattern deals with the challenge of “being locked into proprietary infrastructures, the costs of losing control of all the municipal and citizen data.” This warning echoes a similar one by Sommer Mathis and Alexandra Kanik in City Monitor that claims you’ll be hearing a lot less about ‘smart cities’ due to the growing backlash against big technology companies.”

The disease is properly diagnosed but the suggested cure is odd. City procurement practices are badly outdated. They take too long in an environment where technology cycles are quicker than purchasing decisions.  But more importantly procurement today, in most places, is deeply rooted in process compliance rather than the evaluation of innovation and value added. The answer to the disease is not to refuse any type of proprietary medicine. Risk is an insufficient argument against modernization. And those risks of proprietary technology lock-in or loss of control of data is cured in a contract crafted with the assistance of good legal and technical advisors and back to our definition of the right set of goals. Pointing out this potential issue is great; acting as if it is unsolvable helps no one.

Finally, there are two people and human resources arguments. Saxe, in her New York Times piece, says smart cities will “require a brand-new municipal bureaucracy staffed by tech, data-science and machine-learning experts. Cities will either need to raise the funds required to pay a tech staff or outsource much of their smart city to private companies.”

To which I would return to the definition. Smart cities have smart employees. Contemplating two workforces—one tech literate and the other not—misses the point. There are already millions of city employees who are eager to incorporate new technology and data practices into their work, and to suggest like they don’t wish to perform their jobs – with new tech, data-science, and machine-learning skills – is erroneous at best, and dismissive at worst. Converting some of the outdated job titles to new ones and upskilling existing employees makes sense. Like the 25 chief data officers who meet at Harvard Kennedy School—a job category that largely didn’t exist even 6 years ago – or the internal training academies in Denver and San Francisco that generated large demand from eager employees. Finally, the concentration on expense does not include any sort of return-on-investment analysis for the benefits.

Hans Neubert, a principle at the large design firm Gansler, opined in his criticism that technology might drive us to solve the wrong problems and that we should make sure that “the urban citizen receives a tangible benefit from them.” Again this clearly critical point would not seem to connect at all to the headline of the Gensler piece, “Why Smart Cities Might Not Perform as Well as 'Dumb' Ones.” Smart cities, as we suggest in the definitions, use technology to listen and understand more (no squeaky influential wheels getting all the grease). They use new design tools such as 3D imaging to help neighbors understand proposed changes in the built community. They gather opinions on proposed developments from simple texting. This position is not an argument against smart cities but rather an argument that a smart city (or really any city) should not deploy technology and analyze data unless it directly relates to making the city a better place.

“No matter how much data a city has, addressing urban challenges will still require stable long-term financing, good management and effective personnel,” writes Saxe, “If smart data identifies a road that needs paving, it still needs people to show up with asphalt and a steamroller.” This author is in complete agreement. This is another false choice; there is no pie made up of data and pavement, where each slice of data eliminates stores of asphalt. This false choice is demonstrated in Oakland’s Great Pave, where the city managed to use both data and steamrollers to pave streets that had been neglected for years, often in neighborhoods of color. Smart city efforts focused on improving the quality of life are an imperative. To settle for the unfair status quo is, in fact, dumb.

The lack of a clear definition concerning a smart city makes it too easy for advocates to claim victory, for vendors of point technologies to claim their products advance a city, and for critics to generalize their indictments. A smart city constantly focuses on its residents, using data to improve policy and operations for all who live in its borders, but particularly for those long neglected. As problems grow more complex and “wicked” and governments need new approaches, it’s definitely not smart to stay wedded to a status quo that has not produced fair results. With the right commitment and leadership, smart tools can in fact produce hopeful breakthroughs.


This article was prepared with the help of Harvard Kennedy School writer Betsy Gardner.

About the Author

Stephen Goldsmith 

Stephen Goldsmith is the Derek Bok Professor of the Practice of Urban Policy and the Director of the Innovations in American Government Program at Harvard's Kennedy School of Government. His latest book is Growing Fairly: How to Build Opportunity and Equity in Workforce Development.

Email the author.