How can we be assured smart city tech will do more good than harm? It all comes down to governance

After a Green/Social Democrat coalition won control of the Hamburg state parliament in a February 2020 election, the new government, under Mayor Peter Tschentscher, moved quickly to launch an ambitious transportation strategy for a fast-growing urban region of five million people. “We need to change the way mobility is organized in our city,” says Dennis Heinert, a government spokesperson. (The Free and Hanseatic City of Hamburg has long enjoyed state status in Germany.)

The coalition’s goal is striking: 80 per cent of trips within the city will be via transit, walking, cycling or other shared modes by 2030 in order to cut private vehicle use and carbon emissions. The plan calls for better transit service without fare hikes, a major expansion of the cycling network, and a strategy to load up transit hubs (known as “switch points”) with a range of mobility options, such as e-bike rentals, that cover the last mile between transit stations and home or work.

A central feature of the strategy is the concept of an intelligent transportation system (ITS), which uses various smart city technologies to knit all the pieces together. The elements include self-piloted subway trains and autonomous minibuses, and, eventually, a mobility-as-a-service system that allows travellers to book bike or scooter rentals, carpooling trips or ride-shares from a single app.

How will city officials evaluate the components of the system? “Very easy,” replies Heinert. “Will it help our political goal of the transition of mobility?” State officials, he continues, will vet the portfolio of mobility technologies in terms of how they promote safer, greener and more efficient movement within the region. “There is no project within this whole ITS which is not working towards those goals.”

The oversight of Hamburg’s mobility-technology game plan isn’t difficult to discern: an election brought in a sustainability-minded coalition that wants to advance a program that includes a range of technologies, as well a bureaucratic framework for evaluating those systems. The governance, in other words, is highly transparent, and voters will be able to judge the coalition’s success.

“Governance” is a somewhat nebulous term that orbits around the politics of smart city technology, frequently cited but rarely defined with any degree of precision. At the most abstract level, governance is about accountability. How can ordinary people — and the public institutions that act on their behalf — be assured that these emergent technologies deployed in and around cities will do more good than harm? With the dramatic acceleration of pandemic-related service digitization, as well as the continued rapid growth of so-called platform companies like Lime, Airbnb and Uber, that question has taken on even more saliency.

Sometimes, the contours of smart city governance come into sharper focus by their absence. When Sidewalk Labs’ revealed its master innovation and development plan for a derelict piece of Toronto’s waterfront in June 2019, the hefty four-volume document included a range of governance proposals for how this new smart neighbourhood would be managed.

They included several specially created entities — “Open Space Alliance,” “Waterfront Transportation Management Association,” “Urban Data Trust” — with real regulatory power, nebulous financing arrangements and ties to Sidewalk Labs, but unclear relationships to the municipal agencies (i.e., the public) that perform core tasks like waste or transportation management. The proposals drew sharp criticism, with the city and Waterfront Toronto swiftly shooting down any notion that Sidewalk’s experimental community would march to its own drummer.

So what are we talking about when we talk about smart city governance?

One way to think about this problem is to unpack the layers of governance baked into two very common categories of objects in the public realm: bridges and buildings. Like many of the data or digital devices discussed in this series, both can be seen as systems of engineered technologies, reflecting generations of innovation.

Bridges and buildings are designed and constructed by architects, planners, engineers, and contractors who are professionally trained and accredited, as well as legally accountable for the projects they construct. Both are subject to a range of municipal and provincial policies and regulations, from procurement processes to development approvals, zoning bylaws, design codes, budget expenditures, and transportation plans.

These structures typically involve public consultation and political approval. The decision-making processes, in turn, are mostly transparent thanks to routine disclosure policies and access to information laws. Provincial and federal building codes regulate materials and the minimum standards for their assembly. Inspectors monitor construction prior to completion, and then upkeep and structural integrity afterwards. The fact that a building is privately controlled doesn’t exempt owners from most of these governance systems, nor from their obligation to pay taxes and adhere to laws and regulations.

It’s also worth noting that with both bridges and buildings, the interface to the public realm isn’t left to chance. Local bridges, needless to say, are never off-limits to general traffic. As for buildings, municipal regulations dictate their esthetic and logistical relationships to the outside world, even if those connections may be limited by practical necessities such as fences, secured perimeters, and so on. Governance, in other words, is about the wider context as well as accountability.

Smart city technology, clearly, poses very different questions about the nature and form of urban governance. The devices, software and data, in many cases, are neither tangible nor easily understood. One system may have the potential to do many tasks, some of which have yet to be determined. Data, in turn, is amorphous, fast-moving and malleable. It may be stored not only outside the city but beyond national borders.

Yet some common themes emerge: an expectation of technical robustness and reliability; the existence of professional standards; and the role of policy and regulation in determining how these systems function, including those that have nothing to do with local government but impact public spaces. Finally, adequate governance entails some measure of public engagement and approval, which confer on such technologies and their creators the social licence to operate.

As Rutgers University law professor and smart city expert Ellen Goodman trenchantly observes in a recent essay, cities face a crossroads in their “embrace of the internet of Things and ‘smart city’ agendas. Will they do it in ways that give control over city functions and citizen information to private companies and impenetrable algorithms or will there be public control and accountability?”

Many experts point to Barcelona as a model for progressive smart city governance that balances the Catalonian capital’s desire to attract tech investment with other goals, like citizen engagement, privacy and sustainable development. Barcelona officials began talking about smart city tech in 2011, and the municipal council in 2016 adopted a sweeping “Digital City Plan” designed to ensure that all public services are provided through digital channels. The strategy established a specialized smart city directorate within the municipal government, funding schemes, citizen engagement processes, and a range of technical policies aimed at procurement, data standards and network architecture.

But Barcelona’s outlook is grounded in values, according to Josep-Ramon Ferrer, the former deputy chief information officer and director of the program. Chief among these is preparing the city for rapid 21st-century urbanization while recognizing technology “as a facilitator, not a goal in itself.”

Other cities, of course, have adopted these kind of high-level governance visions, and a growing number have also signed on to even broader pan-urban efforts to ground rapid technology deployment in ethical or humanitarian principals. The Cities Coalition for Digital Rights, launched in 2018 by Barcelona, Amsterdam and New York, includes metropolitan areas around the world that have signed on to a declaration calling for improved privacy policies, more accessible internet access, and measures to ensure that residents have the ability to question artificial intelligence or automated decision-making-based systems to ensure they don’t discriminate or perpetuate hidden data biases.

The City of Toronto belongs to this coalition (as of June 2019), and has begun advancing its own smart/digital city governance policies, largely in response to the controversy generated by Sidewalk Labs. Municipal officials have been hammering out a Digital Infrastructure Plan (which includes a citizen working group whose meetings I’ve attended). It lays out five core guiding principles, including equity and inclusion; effective local government; social, economic and environmental benefits; privacy and security; and democracy and transparency.

Other GTA municipalities have been working on their own strategies. Mississauga’s smart city master plan, developed in response to a federally sponsored “smart city challenge” competition, includes elements such as “living labs” for showcasing new technologies, parks with free Wi-Fi, smart LED street lights, air quality sensors, and the deployment of an AI-powered chatbot on the city’s 311 website. “It’s my job to make them stick,” says Anthea Foyer, a sculptor and former digital curator who is now Mississauga’s smart city program lead.

She cites one proposed initiative that’s bobbed to the surface: the deployment of outdoor digital touch-screens with parks and recreation listings or other municipal information. The devices, however, come with an additional feature: a built-in facial recognition camera that scans eye movement to determine if users appear to understand the content and how the screens function. Foyer has found herself talking to plenty of vendors. “It feels like such a game-changer for me,” she says, noting that the inner-workings of such technologies are “esoteric and hard to understand.” “I would want to make sure residents feel comfortable with it.”

The crux of the smart-city governance riddle, in fact, has to do with what happens between the lofty vision statements adopted by municipal councils like Barcelona’s and the day-to-day choices made by civil servants like Anthea Foyer.

To get to the point where smart city technology is subject to the type of robust governance that applies to buildings and bridges, all orders of governments, but especially provincial, need to move forward with a range of policy, legal and regulatory reforms that will allow cities to make these investments safely. Such changes, moreover, should also apply to technology vendors and platform companies that provide systems or digital services that impact urban regions.

Here’s a partial list of what’s required.

Privacy Legislation Reform

Canada’s privacy laws, as many critics of the Sidewalk Labs proposal observed, aren’t equipped to respond to many of the data-gathering and surveillance technologies that fall under the smart city rubric. Some have been deployed in quasi-public spaces, like malls. Earlier this year, for instance, Canada’s privacy commissioner slammed Cadillac Fairview, which owns shopping centres across the country, for installing concealed video analytics and mobile device tracking systems, and then collecting personal information on patrons without notifying them or seeking their consent.

The federal Liberals recently tabled reforms to national privacy laws that provide individuals with more rights over personal information gathered by private firms and is pressing ahead with implementing a “Digital Charter,” which is, in large measure, a framework for enabling growth in Canada’s information tech sector.

But information technologies such as artificial intelligence tend to advance far faster than public policy, observes Markus Dubber, a University of Toronto law professor and director of the Centre For Ethics.

Case in point: even though the European Union’s General Data Protection Regulation is seen as the world’s most expansive privacy law, revelations from earlier this year about facial recognition apps prompted calls for additional regulations governing these technologies. In fact, even before the Clearview controversy broke, a 2019 analysis by Orla Lynskey, a London School of Economics law professor, warned that “the protection offered by [the EU] legal framework to those impacted by predictive policing technologies is, at best, precarious.”

Regulatory Procedures

Just as significant public sector projects, from transit lines to new gas distribution networks, are subject to environmental assessments that include public hearings, it seems reasonable to expect that significant smart city technology undertakings be subjected to similar scrutiny. These could include requirements that municipalities undertake privacy impact assessments, which are evaluation procedures used in other parts of government, as well as versions of the new federal directive on automated decision-making systems if the smart city system uses AI or machine learning software.

A critical element of this kind of oversight involves testing the durability of so-called “data anonymization” measures. Municipalities release ever-larger tranches of digital information through open data portals, and it is standard practice that any personally identifying information is stripped away.

But a study published last year in Nature Communications itemized many examples of successful de-anonymizing efforts that cross-reference multiple data sets in order to identify individuals in databases considered to be shorn of personally identifying information. (An April 2019, New York Times investigation came to similar conclusions by identifying pedestrians caught on a CCTV walking through Bryant Park in Manhattan using internet searches.) The findings, the authors of the Nature paper note, “question whether current de-identification practices satisfy the anonymization standards of modern data protection laws” in jurisdictions such as the EU and California.

Data Governance, Ownership and Standards

In June 2020, the City of Toronto published an 85-page consultant’s report on “data governance and digital infrastructure,” prepared by a Montreal-based tech policy research non-profit called Open North. A detailed and far-ranging assessment, the document offers what amounts to a 360-degree survey of the largely unresolved policy, legal and tech management issues facing Toronto as it undertakes the kind of transformation cities such as Barcelona have pursued.

It’s a long list that covers everything from approaches to the ownership of data (a hot button topic while Sidewalk Lab’s plans were on the table), ethical uses, gaps in federal legislation and technical standards. While the smart city industry has been roaring along for years, the authors offer a caution: “[D]ata governance in the smart city context is still an emerging field. Therefore, tracking and measuring the outcomes of specific initiatives will require future research.”

A case in point: the twin minefields of “data interoperability” and “open standards.” Such jargon is enough to turn off anyone who’s not a technophile. Yet the principle is as simple as the existence of standardized electrical outlets. There are long-established technical norms for circuits, which means that when you buy a toaster, you don’t have to worry about whether you can plug it into the wall. These concepts apply to some kinds of software, databases, networks and other types of digital infrastructure. They are viewed by many open cities advocates as the means of ensuring that huge technology companies can’t elbow aside rivals by engineering systems that only they can expand — the so-called “vendor lock-in” problem.

Not everyone agrees with these principles, however. Technology critic Brett Frischmann, a professor of law, business and economics at Villanova University in Pennsylvania, questions the need for digital seamlessness because it tends to encourage governments to “over-collect” data in anticipation of future applications. “There’s absolutely no reason to think that the thing that made the internet so successful is what we need for smart cities,” says Frischmann, co-author of Re-engineering Humanity, an examination of the risks of predictive analytics.

For government officials like Anthea Foyer, but also city-dwellers, the importance of open standards in the context of smart city governance is that they prevent technology suppliers, especially very large multi-nationals, from making themselves indispensable (and therefore entrenched monopolies) because no other company’s systems or software can be added on to existing ones.

(Sidewalk Labs attempted just that with a subtle proposal to deploy outdoor mounts dubbed “Koalas” into which its public space sensors could be plugged. These were ostensibly designed to make it easy to upgrade equipment, but the devices are proprietary to Sidewalk/Google, and trademarked, instead of standard USB ports.)

Mark Fox, a U of T professor of computer science, is leading an effort to establish common standards for “city data” through the International Standards Organization. It is a work in progress, he says, that tends to be overlooked in smart city debates because the subject is seen as dauntingly technical. “The adoption of standards is a governance dimension that has received little or no attention in the media, yet it represents the Achilles heel on the path to smart cities.”

Institutional Capacity

Cities employ engineers, planners, architects, public health experts and a range of other professionals who have the expertise to devise and evaluate policies, deliver services, and provide technical input on procurement.

Municipalities, like other large government and private sector organizations, also employ IT staff — programmers, systems engineers, cyber security experts, etc. Yet if local governments intend to invest in smart city technology and infrastructure, they must also be recruiting professionals from disciplines like data analytics, data science, artificial intelligence, data visualization and digital anthropology, all with the goal of creating the kind of bench strength found in other city departments.

The City of Toronto since 2015 has had a big data innovation team, which is primarily focused on transportation applications. But smart city tech cuts across many other departments, so it will be important for municipal officials to ensure that these skills are present throughout the organization.

Beyond the technical aspects, the rapidly expanding role of digital infrastructure and artificial intelligence-driven software demands new approaches to policy-making, especially within municipal divisions whose officials traditionally didn’t pay much attention to technology and data — among them human resources, a field where AI-powered applicant screening tools are increasingly common.

As Toronto’s director of strategy and program management Grant Coffey says of the yet-to-be completed Digital Infrastructure Plan, “This is the first time we’re doing something like this in Toronto.”

Communication

Earlier this fall, Julia Stoyanovich, a New York University assistant professor of computer science, engineering, and data science, and Falaah Arif Khan, a research fellow and artist-in-residence at NYU’s Centre for Responsible AI, published “Mirror, Mirror,” the first of a series of “scientific comics” entitled “Data, Responsibly.”

Although AI might not seem like an obvious topic for a graphic novel, Khan and Stoyanovich (who sat on New York’s Automated Decision-Making Systems task force) have a clear-eyed view of their project. Their aim is to use relatable metaphors to explain AI (e.g., either rule-based recipes or cooking by trial and error) in order to make the concepts accessible to people who don’t have college degrees or deal with disabilities that tend to exclude them from accessing technology or other facets of urban life. “This is the population that is most likely to be hurt by AI and algorithms,” says Stoyanovich, who points to New York City’s recent attempt to regulate employers’ use of AI-enabled screening software in hiring practices. The idea isn’t to side with the “techno-optimists” or with the “techno-bashers,” she adds. “Our goal is really to create a nuanced understanding.”

At U of T, meanwhile, the Centre for Ethics has been hosting multidisciplinary and open-ended public sessions about the applications and implications of the use of AI. As with “Data, Responsibly,” the goal is to yank the subject out of the hands of computer scientists. “[AI] is not a narrow technology-specific issue that should be defined and solved by technical people,” says U of T ethics expert Markus Dubber, who organizes these dialogues. “The more people who participate from different backgrounds, the more they realize there’s no single answer.”

While neither of these projects explicitly target smart city tech, the overlaps are substantial as AI becomes increasingly integral to a wide array of digital and data-driven systems, including those used in AVs, traffic control and policing. Both examples also serve as a prompt for municipal officials to find innovative approaches to citizen engagement.

Consultation on matters such as planning is deeply embedded in our civic culture. But the city’s long-established outreach practices can be rote, exclusionary, inconvenient or just dauntingly bureaucratic. Yet, as Stoyanovich makes clear, the power of these technologies demands, if anything, a far higher degree of public engagement to head off unintended consequences never envisioned in the slick and upbeat presentations of technology companies.

Lastly, communication is also about timely and robust disclosure. Goodman cites the example of Oakland city council, which passed an ordinance in 2018 requiring the municipality to publish detailed annual surveillance reports as well as “surveillance impact assessments” prior to the acquisition of any technology that gathers data that might be used by police. The bylaw also requires the city to seek approval for such investments from Oakland’s Privacy Advisory Commission, which holds public sessions. “Push transparency,” she says, “is really important.”

Democratic Accountability

While researching this series, I interviewed a Dutch academic, Albert Meijer, who has published extensively about smart city technology, data infrastructure, digital governance and other related topics. Despite the Netherland’s pragmatic and upbeat outlook on smart cities, his research has turned up mixed results.

Meijer has developed a systematic way of assessing the success of such investments. He has concluded that there isn’t much evidence that smart city technologies generate value for money — an intriguing result, given the size of the smart city tech sector. Smaller, more focused systems can deliver results, he says, but the more ambitious ones have a way of falling short. “It is technology looking for a problem rather than the other way around,” he says.

In a pointed assessment published in 2019 in the Journal of Urban Technology, Meijer and three other Utrecht University scholars turned their attention to “smart governance,” which they describe as urban governments set up to draw on citizen participation and various communications technologies — online public meetings, social media, software tools, etc. — to make policy decisions.

Despite the proliferation of digital communications channels available to anyone with a smart phone or a Wi-Fi connection, Meijer and his colleagues found that many residents still preferred to engage in person, while those who participated remotely tended to drop out or lose the plot. “The wide net of online activities of many people breeds shallow attention … and transitory involvement,” they observed. “Our review demonstrates that there is certainly no reason for having blind faith in smart governance.”

The takeaway is clear. Cutting-edge digital infrastructure can play a role, either positive or negative, in determining how 21st-century urban regions evolve. But as has been true for millennia, cities will remain defiantly social spaces, populated by humans messily, and often sub-optimally, going about their business.

Truly smart smart cities will embrace this most pedestrian truth of urban life.