This article belongs to the debate » 9/11 und die Überwachung im öffentlichen Raum
17 December 2021

Cities of God

Smart Cities and Surveillance

Set in a lawless, poverty-ridden favela, City of God is one of the most iconic movies of Brazilian cinema. This film is neither about smart cities nor surveillance enabled by digital technology. Also, it is only a faint reference to St Augustine’s seminal work with the same title. Indeed, the movie City of God takes place in the 1970s and 1980s and does not delve into digital transformations. Rather, it is a narrative of an urban space which embodies the experiences of its inhabitants: violence, inequality, cruelty, and poverty. The city is inherently defined by the power imbalances between private actors and between public and private actors. These imbalances become the city and the city becomes these imbalances.

Watching City of God as an adult reminded me of the different journalistic accounts of violence in Brazilian favelas that I, as a curious child, surreptitiously watched or heard through doors left ajar. In particular, it reminded me of the well-known massacre of Candelaria (1993) where multiple children were murdered by members of the Brazilian military police force. As a child, I remember feeling terrorized by the idea that police officers could have killed homeless children sleeping at the doorsteps of a church for what then seemed to be unclear reasons. Much remains unknown about that massacre but many have argued in the meanwhile that it was a targeted killing related to drug dealing which involved a number of highly ranked police officers. Many others felt little compassion with the victims and saw it as a “necessary social cleansing.” And yet others, regarded it as a desolate example of the abandonment of poor black Brazilians. Not surprisingly, only three out of seven accused perpetrators were convicted but released in the meanwhile. Even nowadays, the statistics are cruel to the life expectancy of poor Black Brazilian teenagers.1) They live fast, they die young. This often happens at the mercy of police bullets.

As a scholar, the thought of these violent images invited me to reflect upon what would have happened if the reality portrayed in these accounts of urban inequality had been meticulously recorded by multiple interconnected sensors (e.g., cameras, microphones) and then analyzed using big data, AI, machine learning, and biometrics. Would we have found out who all the perpetrators were? And why they killed children as young as six? Or would surveillance images have been used by the police to better target the victims? In 2019, the student protests in Hong Kong against the use of biometrics gave a partial answer to these questions: Smart-city surveillance is not always used “for the good.” Instead, the faces of regime opponents or, in other contexts, underrepresented minorities, are often self-incriminating elements. Much has been written on smart-city surveillance, biometrics, and their fundamental rights violations, so this reflection may appear at first far-fetched. It is clear that smart cities pose important problems to privacy and that technology-infused urban spaces bring as many benefits as challenges. Nevertheless, the idea of converting underprivileged areas such as favelas into smart cities which has been around for a number of years, is even more problematic and deserves our attention.

This contribution focuses on the attractive force of smart-city technologies for poor and crime-ridden urban spaces and its legal concerns. I argue that we should be particularly critical of the employment of surveillance technologies in slums because they are by definition vulnerable places from different perspectives. First, favelas (and similar neighborhoods) are densely populated areas, with deficient housing where residents have close social interactions. Privacy is limited when large families share limited spaces built close to each other. Privacy becomes impossible when cameras and other sensors are installed in these areas. Second, their residents tend to be vulnerable populations due to their limited income, education, ethnicity, and poor digital literacy. Third, these citizens may not have any bargaining power with public authorities regarding the procurement of these technologies due to their socioeconomic status, lack of mutual trust, and in the specific case of many Brazilian favelas, frequent conflicts between armed militia and the military police. But as this contribution contends, underprivileged neighborhoods as the ones featured in the movie City of God do not need to remain spaces of violence and power imbalances. Also they can become smarter urban areas, if we take into account their characteristics and citizens’ needs.

Smart Cities?

Smart city is an elusive term which can be defined from multiple perspectives. The ‘smartness’ of a city can be for example established by reference to its innovative character, ability to attract human capital, and promote sustainability. In the literature, cities tend to be labelled as ‘smart’ when they promote entrepreneurial, progressive or inclusive policies and make optimal use of available interconnected information to control their operations and manage scarce urban resources. Traditionally, smart cities were characterized by the way in which they harnessed digital technology to improve urban centers. In a smart city, different technologies (e.g., Internet-of-Things, cloud computing, big data) are integrated to ensure that all relevant movements and anomalies are detected and prevented on time. Smart cities gather and integrate real-time information about traffic, crime, water, energy, waste, and ensure that urban services and citizens’ behaviors are adjusted accordingly.

While some years ago, smart cities were regarded as a fad, nowadays it is clear that the industry of smart cities is growing exponentially, including beyond urban centers, giving rise to smart villages. The smart-city industry is largely dominated by Big Tech companies such as IBM and Cisco and it is estimated to be worth $256 billion by 2026. But what exactly does this industry trade beyond cameras, microphones, and other sophisticated sensors? This industry does not sell merely technology. It sells a narrative of modern, efficient urban spaces powered by technology where citizens, public and private actors interact closely to improve city services and spaces. It is a narrative that is told through seductive buzzwords such as interconnectedness, smartness, innovation, safety, health, cultural vibrancy, environmental sustainability, and happiness. Smart cities are nowadays not only technology-infused spaces but also areas of co-creation, experimentation, and close interaction between private actors and public values. The active involvement of citizens has become a key aspect in the development of smart cities. As the governance of innovation processes has become increasingly focused on socio-technical futures, digital innovation and science should be more inclusive and democratic. In smart cities, citizen participation is fostered through different mechanisms such as citizen councils, hackathons, and ‘FabLabs.’ However, the role of citizens as active co-creators of smart-city services is highly disputed and citizens from underprivileged backgrounds are rarely involved. Recent literature has criticized socio-technological systems of governance that monitor citizen power, seek to build trust but are in fact top-down, and its participatory instruments which are often designed for the involvement of tech-savvy citizens. The smart-city narrative is thus a paradoxical tale of digital transformations. The smart-city narrative is designed somewhere between Silicon Valley and the local FabLab, not in the favela. In other words, the smart city is a high-income, highly educated, middle class White narrative.

Smart Favelas

Despite the many paradoxes of smart cities, it has become clear over the last decade that smart cities are no longer a prerogative of developed Western countries or a feature of many Chinese urban centers with ubiquitous surveillance. Instead, emergent economies such as Brazil and India are currently investing in the development of smart cities which can advance their economic growth and promote sustainability. Despite some drawbacks during the lockdown period, examples of smart cities in the mentioned countries include Curitiba and Porto Alegre in Brazil and Pune and Bhubaneswar in India. In Brazil, the use of smart-city technology for predictive policing in Rio de Janeiro is well known. However, what is less known outside the country is the attempt to apply the smart-city narrative to favelas (‘favelas inteligentes or ‘smart slums’). According to the Brazilian Institute of Geography and Statistics, approximately 11.4 million people (6% of the country’s population) live in one of the 6,329 favelas or slums, distributed across Brazil. This sizeable number of people and their interactions have drawn the attention of urban planners and smart city evangelists. However, how should we adapt the smart-city narrative to the reality and needs of favela residents? Sustainable solutions that improve citizens’ access to better waste management, safer water, and better urban planning are surely positive developments. Nevertheless, smart cities have also been long sources of surveillance and privacy concerns. Traditional cities were spaces of amnesia where citizens could move around without leaving any traces of their passage. Smart cities are spaces of memory where every movement is recorded and processed.

As smart cities enter the favela (and other slums elsewhere around the world), we should ask ourselves whether it is desirable to convert these spaces which concentrate the poorest segments of the population into spaces of surveillance and anomaly detection? This issue has already been raised with regards to predictive policing in areas with high-criminality (for example, in New York City). In addition, many of the issues signaled in those contexts (e.g., racial bias, stigmatization of residents) apply to ‘smart favelas’ using surveillance technology. However, when compared to problematic neighborhoods in the Northern hemisphere, favelas and other slums in emerging economies and developing countries are characterized by even more precarious housing, greater violence, isolation from the rest of the city, and limited trust between citizens and public authorities because of police corruption, police violence, and recurrent conflicts.

Smart Cities as Surveilling Cities?

Datafication is a central feature of any smart city, regardless of the perspective we adopt to define and design it. With the deluge of data come multiple privacy concerns regarding the protection of personal and confidential data; bodily privacy; territorial privacy (e.g., personal space, objects, and property); locational and movement privacy (e.g., tracking of spatial behavior); communications privacy (e.g., surveillance of conversations and correspondence); and transactions privacy (e.g., monitoring of queries/searches, purchases, and other exchanges). These concerns arise for example because smart cities place sensors in urban furniture, ranging from waste bins to lampposts, they track phone identifiers, and smart mobility systems rely on extensive geolocation systems. Not all the data collected in smart cities will be qualified as personal data as much of it refers to crowd management, urban or environmental data (e.g., levels of air pollution, traffic density, water level). Also, part of the collected data will become available in the form of open data enabling multiple urban improvements. However, contrary to popular belief, smart cities do not only collect data on the city and its furniture (e.g., lamp polls) but also on their citizens. This data collection has a disproportionately negative effect on low-income underrepresented citizens who are regarded as being at higher risk of committing tax or social welfare fraud or a crime.

While technologies employed for, for example, predictive policing are often overrated as empirical evidence shows their limited effectiveness, their surveillance effect can have chilling and stigmatizing effects on underprivileged populations. Citizens who are subject to surveillance may have lower levels of digital literacy and may not be aware of the extent to which they are being surveilled. At the same time, their low income, ethnicity, criminal record (or that of their family members), and socioeconomic background may make them a target for additional surveillance.

Smart cities and smart favelas do not need to become ‘surveillance cities.’ However, smart cities are difficult to design and regulate so as to fit within traditional public values. The reasons are fourfold. First, the legal category of ‘city’ is almost absent from national legal systems and doctrinal studies (interestingly, the right to the city is a constitutional right in Brazil) and it is highly underestimated in EU law. Cities are thus at different legal crossroads and are meeting points of supranational, national, and local regulations. Smart cities are even more complex to regulate as they are hybrid legal entities where private power can overrule public values. The ‘smart city’ is in most cases a contract or a network of contracts between multiple stakeholders. In other words, a smart city is the result of public-private partnerships which bring together one or multiple municipalities. The contracts and policy guiding smart cities are complex, opaque, and not always aligned with public values. Second, the narrative of smart cities is often top-down, inconsistent with the country’s overall level of development and its citizens’ needs. Third, the protection of fundamental rights is regularly overlooked in the context of smart cities on different grounds. For example, smart cities sell the narrative that data collection is focused on open data and the optimization of services through it. However, allegedly anonymized data pertaining to cities, its residents, commuters, and tourists can be re-identified. Also, there is the risk that personal data collected for specific municipal purposes can be used to nudge citizens to behave differently. Fourth, there is no ‘one-size-fits-all’ approach to smart cities and this applies in particular to underprivileged neighborhoods which require a customized approach to digital innovation. In places such as favelas with deficient urban planning, complex socioeconomic problems, and limited privacy, smart-city technology has to be tailored to these neighborhoods in order to protect the privacy and human dignity of citizens. If this does not happen, digital technology will contribute to deepening already existing problems and power imbalances.

‘Urban smartness’ should not be restricted to cities which are able to afford Big Tech products and services and follow their vision of what a smart city should be. For the urban centers that can afford these technologies, it remains important to inquire whether purchasing digital technology with a neoliberal smart-city narrative in mind is the correct choice. Digital technology can be easily converted into millions of eyes surveilling the most vulnerable members of our societies. For some, smart-city surveillance can be used particularly to watch those who contribute the least to the narrative of smart city and are more prone to be categorized as “the anomaly.” But this doesn’t need to be the fate of ‘smart cities’ and this does not mean either that those underprivileged areas should be closed to digital innovation. On the contrary, smart cities can inspire development at many levels and for different groups. High tech accompanied by surveillance is not the only path. As some African cities already show, urban improvement can also be done through ‘frugal-innovation’ models, favoring do-it-yourself and low-tech solutions, and providing access to new technologies with low investment costs.


At first sight, smart-city projects designed for underprivileged areas may seem appealing as they promise to convert an earthly problematic city (a ‘A City of Man’ in Saint Augustine’s terms) into a ‘City of God.’ However, this conversion (even if possible) may come at a high price for fundamental rights and public values if we allow for a top-down, Western, and Big Tech narrative to be implemented. This is a narrative where few speak and those who do are rarely embedded in the local community. Surveillance technologies for smart favelas or other underprivileged neighborhoods are thus designed primarily to detect ‘anomalies’ thus deepening existing power imbalances. Nevertheless, in top-down smart cities, the children sleeping at the doorsteps of the Candelaria church are the anomaly. The teenagers that die before their time in the movie City of God are the anomaly. The poor are the anomaly. But it doesn’t have to be this way.

We return here to where we started: the Brazilian movie City of God where technology (in the form of a simple camera) and the recording of information are symbiotically connected to the city. For those who have not watched the movie, the main character is a photographer reporting for the local newspaper. The information he records through the lens of his camera is an accurate representation of the favela because he is socially embedded in it. Without these social and human connections, the recording and processing of information about a violent favela could have been used to deepen existing stigmas and conflicts, thus turning it into more of ‘a City of Man’ rather than ‘a City of God’. Smart cities can be surveillance spaces detached from the real citizenry and its needs. But they can also be ‘cities of God’ where digital transformations are co-created with communities to address power imbalances, where there is a place for frugal innovation that fits locals’ needs rather than sophisticated systems designed for establishing risks and predicting anomalies, and where different groups—regardless of their income, literacy, and skin color—always have a say.

Many thanks to Francisco Abreu Duarte, Catalina Goanta, Luisa Netto, and Moritz Schramm for their insightful comments.


1 Of the 56,000 homicides in Brazil every year, 30,000 are young people aged 15 to 29. 77% of them are black,
see Amnesty International, Young, Black, Alive: Breaking the Silence on Brazil’s Soaring Youth Homicide Rate, Amnesty (November26, 2014), at (last accessed on December 12, 2021); for a more recent account, see The Economist, Rio de Janeiro Asks Why Its Cops Kill So Many Black People, The Economist (August 14, 2021) at