Smart cities, smart borders: Sensing networks and security in the urban space
In the outskirts of Kolkata, West Bengal, a satellite township called Rajarhat New Town is being transformed into a smart city, as part of the ‘100 Smart Cities’ programme launched by the Indian government in 2015. The township was originally designed, about thirty years ago, as a Special Economic Zone (SEZ) for the IT industry but resulted into a paradoxical space where corporate enclaves and slums, upscale hotels and unfinished constructions uneasily coexist. The projects for New Town reiterate the narrative, crafted by major commercial players, of smart cities as smoothly interconnected systems, and promise that the extensive distribution of computing technologies will turn this urban purgatory into an efficient and harmonic environment. This chapter deconstructs this storyline and draws attention to the ways in which processes of digitalization entail the distribution of border technologies across the urban space. I also discuss how these bordering processes might be constitutive of distinct politics of knowledge and aesthetic, as well as of new techniques of security and urban government.
In her work on the introduction of biometric borders in the context of post 9/11 ‘war on terror’, Louise Amoore (2006) explains how these become ubiquitous and bring risk profiling techniques into every realm of social life. Smart borders are informed by an anticipatory logic that seeks to identify, assess and authorize (or not) individuals in such a way that ‘the body itself is inscribed with, and demarcates, a continual crossing of multiple encoded borders – social, legal, gendered, racialized and so on’ (2006: 337). More recently, Holger Pötzsch (2015) has described the emergence of a sociotechnical apparatus – what he calls the ‘iBorder’ – made of biometrics, dataveillance and AI, which generates
bordering processes that disperse locally as well as across transnational space. In these processes, individuals become objects of governance to be analysed and assessed, but also serve as implicit contributors to the database enabling algorithm-driven mappings of patterns of behaviour and association. (Pötzsch 2015: 23)
In the past few years, studies on the introduction of smart borders have explored how digital technologies and algorithmic calculations are transforming security practices and responses to terrorism and migration movements in Europe and North America (de Goede et al. 2014; Leese 2016). At the same time, scholars have noted that smart borders are increasingly seeping into the city and neighbourhoods (Amoore 2006; Amoore, Marmura and Salter 2008) as part of new military and security paradigms, emerging in the US and UK , which problematize urban life (Graham 2012). However, work remains to be done to chart the specific, situated ways in which smart borders permeate and constitute urban environments, especially in cities outside the US and UK, where the category of military urbanism might not be equally relevant.
At the same time, critiques of smart cities abound, and point to the risks of technocratic governance, surveillance, perpetuation of inequality, or social engineering (Crang and Graham 2007; Halpern et al. 2013; Kitchin and Perng 2016). Again, Stephen Graham (2012) has pointed to the ways in which the digitalization of urban life spreads and normalizes technologies that were developed for military purposes. Overall, this critical literature has hardly ever crossed over to a more punctual and comprehensive discussion on borders in smart cities. Borders have a polysemic, heterogeneous and dynamic nature (Balibar 2002; Mezzadra and Neilson 2013). They work along, within and beyond the territorial limits of state as instruments of differential inclusion, as well as exclusion, which continuously filter and stratify the circulation of people and things. This chapter illustrates how, by creating a connected and sentient environment (Crang and Graham 2014; Thrift 2014; Gabrys 2016), digital infrastructures also perform and distribute border functions across the urban space.
In the making of smart cities, Rob Kitchin and Sung-Yueh Perng (2016) note, code becomes embedded in urban infrastructures, services and utilities, and government practices, in modalities that are always contingent and situated. Cities under digitalization can be seen as a patchwork of millions sociotechnical assemblages where code is, at once, produced through and producing multiple sets of relations with other material and discursive elements (Kitchin and Perng 2016; Dourish 2016). Empirical researches confirm how diverse and complex these relations can be. For example, Ayona Datta (2017) observes how the strategies to forge new smart citizens in the wake of India’s 100 smart cities challenge merge a global imaginary of smart citizenship with the issues and struggles of postcolonial citizenship, resulting in hybrid and vernacular forms of digital engagement in Indian cities. In his work on data-driven urbanism in Delhi, Sandeep Mertia (2017) illustrates how the forms of knowledge production, forms of authority and identities in and about the city are being reconfigured through sensing/computing infrastructures in ways that are contingent and very much affected by contextual factors. The sociotechnical assemblages that compose a smart city have a political significance that demands attention. For this reason, I look at the frictions and barriers that take place around and through these assemblages from the angle of borders. The point here is neither to fetishize the notion of borders, nor to offer a fixed spatial representation of instrumented cities. Rather, looking at urban digitalization through the lens of borders is a way to attend to the distributed, situated and often microscopic relations of power that permeate smart infrastructures.
This chapter is based on the examination of planning documents, direct observation and interviews with informants involved with the process of urban digitalization at various levels. It is organized as follows: The first section explores how popular narratives of smart cities as harmonic, seamless systems have been crafted through a set of assumptions and topoi, in accordance with specific commercial strategies. The second section reviews the history of smart developments in New Town, and illustrates how digitalization has in fact taken place through zoning processes. In the third section, I examine the dissemination of border techniques across digitalized infrastructures, objects and apps of common use, and how the promises of smart urban harmony actually turn into the multiplication of points of control and filter in every aspect of urban life. The fourth section investigates how sensing and computing systems reconfigure categories of perception and knowledge, as well as relations, by setting boundaries and filters, and how borders are active at an ontogenetic level. In the conclusions, I situate these analyses in a broader perspective, and argue that processes of digital bordering cannot be classified merely as examples of surveillance or dataveillance. I suggest instead to look at them as infrastructures of preemption and anticipatory government.
Smart city narratives
It can be said that Smart cities of the Future will be smoother, more social, and more open than they are today.
Marketing and eMobility Team Leader at Bosch Belgium1
Arrows in vivid colours run between skyscrapers, ports, parks and highways. Footages of people using smartphones and tablets flow quickly. Wall-size dashboards show interactive maps, graphics and figures. Smiling testimonials tell stories of success and profess their faith in a digital future. What I am describing is not a commercial video of smart city solutions released by a major provider. It is, virtually, all of them. IBM’s Smarter Cities, CISCO’s Smart+Connected, Microsoft’s City Next, SAP Future Cities, are only some of the products on the growing market of urban digitalization. And while competing against each other to secure contracts with city governments, these and other corporate players have contributed to forge a model of smart city that is, to a large extent, homogeneous. Their corporate documents and advertising resort to the same imaginary, the same jargon, the same visual style. The key topics in these narratives – efficiency, sustainability, resilience – are perhaps better described as topoi, such is the frequency and the uniformity in which they recur. In all these smart city systems, the focus is on ‘breaking the silos’ between different urban datasets – traffic, waste, pollution, energy, crime, social programmes, healthcare, education etc. – and creating one integrated platform for the analysis of data – a single view of the city. This is achieved by distributing IoT networks across the city, and by running analytics across disparate domains, from sensors and video cameras to social networks.
All these corporate documents present the creation of smart cities as a smooth, harmonic process, based on the assumption that more automation necessarily equals more efficiency, safety and sustainability for all; and that the integration of systems will proceed seamlessly.
Scholars have critically investigated the genesis and evolution of the predominant smart city discourse ad the underpinning commercial strategies. Donald McNeill (2015) demonstrates how the launch of IBM’s Smarter Planet campaign in 2008 has signalled a substantial restructuring of the company, which sold its PC division to Lenovo in 2004 with the intention to concentrate its business in the emerging sector of IT consulting. Having identified cities as a high-potential market, IBM started to focus on aggressively promoting its solutions for urban management. Analysing these commercial strategies, Ola Söderström, Till Paasche and Francisco Klauser (2014) suggest that popular narratives of smart cities can be read as a form of ‘corporate storytelling’. Drawing on the concept of ‘obligatory passage point’ (OPP) proposed by Michel Callon, the authors show how IBM has forged discourses ‘that presents their smart technologies as the only solution for various urban problems and hence becomes an OPP’. (2014: 310). In 2011, the tech colossus officially registered the term ‘smarter cities’ as a trademark, while continuing Smarter Planet’s powerful advertising strategy, including free consultancy for municipalities, international conferences, research papers, videos, and so on. Across these different outlets, the city is presented as a ‘system of systems’ – a theme then adopted by some of IBM major competitors, such as Microsoft and Cisco – and broken down into nine ‘pillars’, which represent the relevant sectors that have to be digitally integrated to optimize urban government. In other words, the city, along with all its issues and components, is translated in the language of data and algorithms (Söderström, Paasche and Klauser 2014: 313). Datafication and automation are associated with a number of beneficial results – transparency, efficiency, cost-effectiveness, inclusiveness, sustainability, safety and so on – up to the point that they become synonyms for better government and liveability. The processes of interconnection of infrastructures, devices, data and management practices are supposed to happen linearly and without frictions, and to be inherently virtuous. It is largely through the articulation between these discursive moves and the considerable economic power of a colossus like IBM, that the mainstream label of smart city has taken shape. As this storyline continues to be echoed among tech companies, consultants, city officers and media, the smart city is uncritically presented as a progressive, and somewhat necessary evolution of the urban condition.
The narratives of smart cities mobilized in New Town Kolkata do not deviate much from the corporate version. On the website of the India Smart Cities Mission – the government programme within which the transformation of New Town is taking place – smart cities are vaguely defined as ‘clean and sustainable environments’, where ‘layers of smartness’ are added on comprehensive infrastructural development (Smart Cities Mission, n.d). The list of technological solutions that make a city smart resembles quite closely the dominant commercial models. The city is broken down into relevant components – administrative services, waste management, energy, water, mobility, health and business – that are supposed to be equipped with digital technology and managed via analytics.
The core idea of adding ‘layers of smartness’ presupposes a linear development process, where technological elements and governmental practices interconnect progressively and without frictions. New Town’s municipal authorities have also perpetuated this narrative throughout activities of dissemination and citizen engagement conducted with the help of consultants, such as British company Future Cities Catapult. In the workshops and events organized for the middle class residents of New Town during 2016, participants were educated about the benefits of upcoming digitalization, and invited to contribute ideas as to how to add more smart solutions to pre-selected areas of intervention – water and energy, transports, security, health, administrative services. The outcome of this ‘participative’ design phase is depicted as a green, harmonic landscape, of which relevant components are provided with sensing technologies and interconnected.
In 2015, New Town Kolkata applied for the Smart Cities Challenge, a competition-based funding scheme launched by the Indian Government with the aim to transform 100 cities into digital and sustainable cities, and worth approximately US$ 15 billion overall. Before that, the development of New Town had progressed quite controversially.2 The township was planned in the early nineties as a Special Economic Zone (SEZ) for the IT industry in the rural area of Rajarhat, in the north-eastern fringes of Kolkata. Strong protests rose as the former ruling Left Front government forcibly expropriated lands from farmers and villagers; thousands faced police brutality, were jailed or killed. In the following years, business parks, gated communities and luxury shopping malls began to rise alongside wastelands, villages and slums. Many of the dispossessed farmers remained in the area, living in informal settlements and taking up precarious, low-paying jobs as domestic workers, security guards, street vendors. Largely driven by speculation, the development of New Town was hampered by the financial crisis of 2008, resulting in a paradoxical urbanscape of unfinished infrastructures, unsold houses, highly securitized enclaves and stray cattle. In 2011, Ananya Roy described the township as ‘the ghost town of homegrown neoliberalism, one where the ruins of the suburban middle-class dream are starkly visible’ (Roy 2011: 275). Attracted by the low cost of labour and lands, several IT firms such as IBM, Tata Consultancy Services, Wipro and Accenture established branches in New Town, where they run the more basic and menial tasks of the industry such as software beta testing or business process outsourcing (Rossiter 2016). As New Town seemed to be stuck in a condition of suspended development, and disturbingly veering towards urban dystopia (Dey et al. 2013) the Smart City Challenge likely appeared to local authorities and investors as a chance to resurrect the fortunes of the township.
The Smart City Proposal (SCP) for New Town is not much of a consistent document. Developed through negotiations among several public agencies, consultants and economic stakeholders,3 the proposal revolves around the Pan City Solution, a system of integrated digital infrastructures and software for the management of the city. On one hand, in tune with the standard vision of smart city promoted by IT firms and consultants, the SCP aims to develop a sensing urban environment, where infrastructures – from bus shelters to waste bins, from water meters to light poles – are extensively provided with sensors, GPS trackers and cameras, while several urban services are provided via mobile applications. The data sourced from sensing infrastructures are then integrated, cross-checked and processed via analytics into a single command and control room. At the same time, and quite at odds with its claim for innovation, the plan includes very basic elements of urban development – i.e., sidewalks, public toilets or street lights. Overall, Pan City looks like a sort of vernacular version of mainstream smart city projects, where the effort towards fast digitalization coexists with the need to provide basic infrastructures and services in the area. The contradiction between the aspirations towards a global model of urban development and conditions of widespread poverty, inequality and lack of essential facilities, is crucial to understand how borders intervene in the process of digitalization.
In the first stages of development of New Town, marked by political contestations and social unrest, the implementation of digital technologies took place behind the walls of upscale private developments protected with security checkpoints, biometric identification, x-ray scan and CCTVs. Within the gates of business districts like Ecospace or Tata’s Gitanjali Park, smart infrastructures – high-speed Internet, security software, Building Automations Systems (BAS) that control ventilation, temperature, power systems and water through the IoT – have been running for a few years now. The informal sector is kept out from these enclaves, or only admitted inside as service workforce – cleaners, guards, gardeners. More in general, a large part of the population of New Town still struggles to access the Internet and digital devices. According to the Internet and Mobile Association of India (IAMAI), India has approximately 450 million Internet users (IAMAI 2019), slightly more than one third of the overall population. But while technology is becoming cheaper and definitely accessible for wide strata of the population, smartphones, laptops, computers and Internet connectivity are still out of reach, at least on regular basis, for households and individuals that live in slums and work precariously in the informal sector. Between the smart world of tech companies, and the life of New Town’s urban poor there is a gap of income, education and social agency that persists in the processes of urban digitalization.
At this stage, Pan City is designed as an Area Based Development (ABD). Through a digital citizen polling on the MyGov website, one district of New Town has been selected to be transformed into a smart area, where the new technologies and management systems will first be tested and implemented. The zone identified coincides with Action Areas 1A and 1C, the most densely populated in New Town, the closest to the periphery of Kolkata and to the IT hub of Salt Lake Sector V. In Action Areas 1A and 1C, the implementation of infrastructures is more advanced than in the rest of the township, urbanization appears slightly more consistent, and informal settlements have been largely cleared out. Strategic facilities, like a water treatment plant and the central bus station, are located here, as are some of New Town’s most important business sites and landmarks, such as the NKDA headquarter and the monumental Biswa Bangla Gate. Meanwhile, outside the borders of the designated smart zone, large portions of New Town remain deprived of basic services and infrastructures. In Action Area II, just a few miles away, cutting-edge IT campuses are punctuated by informal markets and bustees where running water and sewerage do not reach. The landscape remains similar in the residential towers of Action Area III, a little further east, where seemingly abandoned building sites and the skeletons of unfinished towers stand out among wastelands. Such entanglements of hyper development and deprivation are far from uncommon in most megacities in the country; in fact, they can be seen as a major feature of Indian urbanization (Schindler 2014). The same applies to the increasing securitization of private and public spaces, over the past two decades, that filters the interactions between different urban worlds, while also introducing new forms of exploitation of the informal labour (Gooptu 2013). So far, at least in New Town, digitalization has not reversed these tendencies, but has rather grafted onto them. Smart developments have largely concentrated within clusters of privilege, and access to them has been restricted on the basis of class and labour control.
This overview illustrates how the making of smart New Town Kolkata is taking place through the formation of hubs and enclaves where digital implementations are concentrated. I refer to this process, which is in sharp contrast with narratives of smart cities as seamless, harmonic environments, as digital zoning. As we learn from a rich body of literature, zoning techniques are always infused with political effects and power relations. Much attention has been paid, for example, to the key role played by the creation of Special Economic Zones (SEZs) and logistical corridors in positioning countries like China and India, and South-East Asia more in general, in the global economy and political relations, as well as in transforming forms of accumulation and extraction, labour relations, normative arrangements, and lifestyles (Ong 2006; Easterling 2008; Mezzadra and Neilson 2013). There are no zones without borders; and zoning processes, be they on a larger or smaller scale, are often the occasion where techniques for monitoring and filtering the movements of people and things are tested or recalibrated. The processes of urban zoning have often been associated with the notions of enclavism (Atkinson and Blandy 2005) or enclave urbanism (Angotti 2013), to describe how the creation of gated, securitized compounds for residential, commercial or leisure purposes increasingly marks neoliberal urban developments and rising inequalities between social groups. Many elements of the development of New Town in recent years, including the creation of gated communities and business parks, can be seen as examples of enclavism. However, this category does not exhaust the complexity of the zoning processes that are associated with the smart city projects. Urban digital zones have emerged in multiple, flexible and informal manners, and have produced multifaceted effects. Some of the zones that I described in this section, such as New Town’s Area Based Development and SEZs, are formally established via legal acts, while others, i.e. corporate enclaves, are demarcated de facto, in informal but no less effective ways, through conspicuous securitization and the restriction of access only to a certain class of citizens. These zoning processes, through which smart infrastructures are being tested and implemented, reflect the patterns of inequality and social hierarchization that have shaped the creation of New Town since the beginning. Rather than connecting the urban environment seamlessly and inclusively, as the smart city narratives promise, the processes of digitalization embed extant socio-spatial borders and produce new ones, which separate and filter the population of New Town along the lines of class and social agency.
Not only are borders traced around digital infrastructures in the making of smart cities; they also become incorporated in a wide range of mundane objects and activities, and therefore ubiquitous across the urban space. The computing systems onto which smart city projects rely are, indeed, built around algorithmic techniques of classification, identification and profiling that are currently in use for the management of national borders, as well as for policing and crime investigation. The smart solutions laid out in the Pan City Solution for New Town disseminate border technologies across every domain of urban administration, from water supply to tax policies, as well as in a number of everyday obvious activities, like getting on a bus or taking out the garbage.
As mentioned in the previous sections, New Town’s Area Based Development (ABD) is supposed to be the first step of the proposed smart city. Not dissimilarly from many other smart city projects, the ABD is designed as a space where ideally every house, vehicle, public area and piece of infrastructure is equipped with sensing devices, connected to the urban network, and managed via a single, central platform. According to New Town Smart City Proposal (2016), the urban components that will be integrated in the digital platform include:
- Air pollution monitoring: sensors for air quality monitoring will be installed on light poles and display real time data on LED display boards in strategic locations of the area;
- Smart parking: nine smart parking areas with parking sensors installed in light poles to collect data from the cars. At least four have been introduced already, in partnership with Indian app Park24x7 – a mobile app that allows users to book in advance and pay for their parking online;4
- Sewerage and drainage monitoring: Sensor-based drainage covers will send signal to the control room on the quantity of rainfall in the area and will activate pumps to avoid water logging. More sensors will be installed to monitor sewerage and drainage and transmit the data to the Pan City control room;
- Project Zero – solid waste management: All waste collection vehicles will be equipped with GPS and tracked by the control room. Sensor-based e-bins will be installed in public areas and tracked through Off-Site Real-Time Monitoring (OSRT);
- Smart metering: All conventional meters for water and electricity will be replaced with smart meters. This will allow remote meter reading, monitoring of load profile, monitoring of tampering/pilferage by consumer from the control room. The water distribution lines will be equipped with Supervisory Control and Data Acquisition (SCADA) system, including sensor-based transducers and flow meters;
- Safety and security: CCTV cameras will be set up on light poles for 24/7 surveillance. Real-time video content analysis will be performed in the control room. 2000 intelligent street lights will be installed as well as panic buttons at key points, connected to the control room for emergency response. Drones will monitor civic services such as road conditions, street lights, littering and waste management;
- Health: Telemedicine kiosks will be installed in every block to deliver primary medical services. Healthcare for residents will be managed via mobile apps and a Smart Watch programme supported by volunteers;
- Mobility: Public vehicles including Electric buses, Autos and Totos will be monitored via GPS from the control room, while information on routes and timetable will be available on a mobile app.
The Pan City Control Centre is where data are gathered and visualized to monitor and manage all the critical components of the smart city in a holistic manner. Once processed via analytics, data turn into models and alerts and are displayed on a central dashboard which provides real-time diagnosis of urban components, from traffic congestion to the quality of the air, from water consumption to garbage disposal. In other words, in the planner’s vision, the entire city becomes incorporated into a system of non-stop monitoring and risk assessment. What is commonly presented as seamless interconnection, efficiency and transparency in fact disseminates the logic and practices of border management across every domain of urban life, often on a microscopic level. Common utilities and ordinary activities become the vectors of techniques of identification, profiling and scoring. Real-time data on power consumption sent from smart meters are automatically crossed with information on housing occupancy and shared with the police, to detect potential ‘illegal’ residents. The network of telemedicine kiosks and health-related apps elaborate profiles on both the individual and collective levels of health or diseases in the city. Mobility apps record the itineraries of people across the city, as well as their use of public transport, cars, taxis, of other vehicles. While light poles and bus stands double as surveillance spots, drones provide bird’s eye monitoring. As most of these projects are still underway – their implementation outsourced to private partners such as Intel, HP, SAP, Oracle, and the like – or even on paper only, it is too early to assess their effects on urban life. But what matters for the sake of this discussion, is that they already present the logic of the future urban environment. In the Pan City Solution, the narrative of a smoothly interconnected city translates into a landscape of ubiquitous borders. Techniques for scrutinizing and filtering are built into every bit of the urban sensing systems. Increasingly, the interactions between the population and the urban infrastructures and services are mediated by digital identification, and feed processes of algorithmic profiling and modelling.
Social media constitute a further domain of monitoring. From the Smart City Proposal, we learn that the city is negotiating with Abzooba, an Indian company specialized in Artificial Intelligence, about installing Xpresso, the company’s proprietary Natural Language Processing (NLP) software, to gather and process data concerning New Town on social media (NKDA 2016: 98). NLP is a specific segment of Artificial Intelligence (AI), which makes it possible for computers to read and understand human language and process large volumes of unstructured data, such as social media content. Xpresso was originally developed to help companies analyse customer feedbacks and improve their commercial strategies accordingly. In the customized version for urban management, Xpresso will help urban authorities exploit large volume of unstructured data, such as social media content, and gain
[…] a structured bird eye view about different aspects (Police, Transportation, Healthcare, Water, Road etc.) of city and citizen sentiment (positive, negative, neutral) about each of these aspects. (NKDA 2016: 98)
The application runs cognitive bots that are able to translate ‘text into context’,5 understand the nuances of human expression and classify the intents of those who write. By generating actionable information, Xpresso provides real-time monitoring as well as an ‘early warning system’ to anticipate potential problems. When high percentages, temporal or spatial spikes in negative sentiment, such as anger or fear, or large number of complaints on selected topics are registered, the dashboard displays specific alerts. Authorities are able then to ‘drill down’ to view complaints in detail, and take ‘corrective measures’ (NKDA 2016: 98). A case study on the Abzooba website describes how Xpresso has been experimented before in the management of urban data.
According to the case study, Xpresso generated several benefits in urban management, including the capability to measure public opinion, make more informed decisions on new policies and better evaluate existing policies; ‘safeguard the country’s reputation’ (sic) by monitoring social media conversations, and how these might affect overseas investors and tourists opinion of the country; anticipate disease outbreaks by correlating searches for specific symptoms, and improve disaster response by understanding the situation on the ground; prevent and mitigate potential crisis through ‘active listening’; and ‘transform security clearance process’ by leveraging social media data for ‘national security, background investigations, programme integrity, insider threat detection, and more’. Of course, Abzooba is not a pioneer in the field. Opinion mining and sentiment analysis are standard methods for the organization of social media content and related commercial strategies. A number of systems are being developed, not only by IT corporations, but also by academic research groups, to perform real-time sentiment analysis of discrete social media streams, that assess, for example, how urgent specific urban issues are perceived by citizens (Masdeval and Veloso 2015); the spatial distribution of intolerant discourses in Italy, or the community feelings about recovery from earthquake in the city of L’Aquila (Musto et al. 2015); or to monitor, more in general, the ‘situation’ of specific urban areas that emerge from topics and emotions on social media (Weiler, Grossniklaus and Scholl 2016).
The adoption of a software like Xpresso is also part, I suggest, of the bordering process that are shaping the making of the smart city. As explained earlier in this chapter, access to digital technologies in New Town remains far from universal. A considerable part of the township’s population is not able to be active on social media on regular basis, or ever. In this context, monitoring the city and its citizens via social media is a form of pre-selection, or differential inclusion (Mezzadra and Neilson 2013) of the data that are relevant to urban government. In other words, only the voices that can be expressed on digital platforms count as urban data (even if for monitoring purposes only); and only those who provide data count as citizens. The example of Xpresso in New Town subverts the usual understanding of dataveillance. While common concerns are about being tracked, spied and manipulated through our immersion in digital technologies, there are groups of people that are not subject to dataveillance because their socio-economic conditions are below even that. Ned Rossiter (2016) uses the term ‘post-population’ to describe those who escape algorithmic control on labour or social life but pay the price for this anonymity or ‘ungovernability’ with extreme precariousness and vulnerable conditions, such as the dispossessed farmers and slum dwellers of Rajarhat. In the making of smart New Town then, social media emerge as the terrain of a twofold filtering process. On one hand, access to social media qualifies citizens. On the other hand, those who count as citizens (in their capacity as data providers) are subject to practices of monitoring and profiling.
The secrecy around the algorithms and code strings that process urban data – from those generated by sensing infrastructures, to social media – can be seen as a further bordering process. In the accessible documents about New Town there is no mention of the analytics settings employed in the software that run city systems, or of the specific pools of data in use. Most likely, this information belong to the software provider, and are therefore protected by corporate cyber-security. Even city officers and agencies that have to authorize interventions and elaborate policies on the basis of analytics have no access to the raw data, or to the algorithmic settings. The ways in which the profit strategies of software providers and consultants might have informed the sourcing and processing of data; or how biases and specific understandings of social and environmental categories can be silently embedded in the calculative framework – all this is withheld from public discussion and critique. Despite promises of transparency and evidence, the operational core of smart urban management remains opaque and hidden underneath layers of digital barriers, protocols and private agreements that come with the application of smart technologies to cities.
A new partition of the sensible: Borders and digital ontogenesis
The previous sections of this chapter have described how urban digitalization proceeds by establishing borders and zones, and by disseminating border techniques – of monitoring, measurement and filter – across infrastructures and devices of common use. But these bordering processes are active also in the sphere of perception, cognition and relations. In her book Program Earth, Jennifer Gabrys (2016) combines the notions of ‘concrescence’ formulated by Alfred North Whitehead and that of ‘concretization’, proposed by Gilbert Simondon, to describe how computing environments come into being. Sensing/computing systems, Gabrys claims, are more than assemblages, more than a mere aggregation of sociotechnical elements. In fact, they are able to generate new relations between elements, new forms of connection, expression, knowledge and actions; they have, in this sense, an ontogenetic quality. The making of computing environments is, therefore, a relational process, where computing becomes environmental while at the same time, the environment becomes computational. Gabrys also draws connections between this understanding of the environment and Foucault’s notion of milieu as the field where security and government operate, and of environmentality ‘as a spatial–material distribution and relationality of power through environments, technologies, and ways of life’ (Gabrys 2016: 187). Hence, focusing on the borders that emerge from the processes of digitalization is a way to grasp how power relations are articulated across sensing/computing environments. As techniques of monitoring, identification, and profiling become embedded into mundane objects and infrastructure, they define a distinct terrain and distinct trajectories of government.
In his book The Politics of Aesthetics (2004), Jacques Rancière argues that any social order is constructed through a specific distribution of the sensible. This concept indicates modes of perception that set the boundaries between what can be seen or not seen, said and not said, heard and not heard, measured and not measured, and ultimately, between what is licit or illicit. Social roles and forms of participation are defined through specific distributions of the sensible which can, at once, include and exclude. In this sense, every social and political system is in the first place an aesthetic regime – where the term ‘aesthetic’ refers to what is experienced through senses – insofar as it is organized through distinct forms of perception and sensorial relations among humans, objects and nature. While Rancière’s own analysis engages in a detailed examination of historical examples of politics of aesthetic, here I appropriate the notion of ‘distribution of the sensible’ and put it at work in a very different context, to analyse how smart technologies are increasingly performing bordering functions and reconfiguring urban life and government. The distribution of the sensible is, I argue, part of the ontogenetic processes discussed by Gabrys (2016), as changing forms of perception shape the ways in which relations unfold between the various environmental components. Looking at the reconfiguration of the senses and at the creation of new modes of existence that connect humans and things is key to understand how the computing milieu is governed.
How do sensors and analytics produce new distributions of the sensible in the city, and with what effects for the human and non-human elements involved? How is this distribution of the sensible relevant to the production of security and urban government?
When sensing technologies – in their various versions: trackers, beacons, cameras, wearables, smartphones and applications – are applied onto urban components, they enable new modalities of perception and interaction. They remodulate the patterns of attention towards the object, resource, or activity concerned. They can invite and even force attention from users, or, conversely, they might deliberately avoid it, when they are invisible. They signal that a certain component is important in the urban system. They warn that what happens around it is going to be scrutinized and assessed. Whether demanding or rejecting attention from humans, sensors are definitely attentive to selected dynamics, and at the same time, indifferent to others. In doing all this, they reconfigure the order of things, perception, thoughts and action. As described earlier in this chapter, this happens through specific techniques of monitoring and identification. Situations that could previously remain unnoticed, such as the number of people crossing the street at a certain junction, the quantity and quality of particle in the air, the amount of garbage in a bin, become, through the application of sensors, necessary points of application of the urban attention. This attention is political and unfolds simultaneously on interrelated levels. First, it demands the engagement of citizens, which are required to take part into the sensing process, by sending data, remaining aware of the information available, and behaving accordingly. At the same time, it also dictates the modalities in which this interaction can take place: through the mediation of digital devices and platforms. Second, while contributing to the monitoring activity, citizens become objects of scrutinization themselves, through the ubiquitous practices of profiling described before. Third, it marks the specific targets of urban policies and intervention: where there are sensors, there is also government. Fourth, as a whole, sensing networks produce a new map and a new definition of what is to be perceived and lived as a urban system.
The distribution of the sensible continues through analytics processes, where the performances of urban components are algorithmically broken down into factors of normality, deviation and risk, and then re-assembled into predictive models. Here again, the work of algorithms sets out distinct boundaries between what can be seen or not seen, made actionable or not. It is important to pay attention to the modalities in which analytics and modelling render urban elements, determining what is worth paying attention to, what is worth measuring. A significant epistemic move is visible here, as the very practice of measuring becomes the measure of worth itself. In other words, if something is not monitored and measured, if it is not inscribed in the computational grid, it has no worth in the smart urban system. In this sense, algorithms create new regimes of visibility and worth, which are politically charged. At the same time, a new regime of invisibility is created, that is the one of code strings and operative systems that process urban data. As noted earlier in this chapter, these crucial components remain largely inaccessible not only to citizens, but also to the city agencies that are expected to act upon data.
To conclude this discussion of the partition of the sensible, I maintain that the ontogenetic power that Gabrys assigns to sensing/computing environments reconfigures the order of the cognitive, aesthetic and relational processes. In other words, borders operate at an ontogenetic level, insofar as the forms of classification and filter that come with extensive datafication are able to reshape the apprehension of reality, and the relations between human and non-human elements. They reconfigure, at once the milieu where security and government operate; and the modalities through which they operate.
Conclusion: Beyond dataveillance
What emerges from the examination of New Town smart projects is an urban landscape where bordering functions – identity verification, biometrics recognition, profiling – are immanent to the development of digital infrastructures. This is evidently in contrast with popular narratives of smart cities as seamless, smoothly interconnected spaces. I have outlined three main dimensions where borders operate. The first considers the processes of digital zoning through which smart technologies are introduced and tested in the urban territory. The second one concerns the fact that practices of identification and filters are pervasively attached to objects, devices and software that are in use for everyday activities. There is, then, an ontogenetic dimension, where forms of measurement and classification enacted by sensing and computing systems are able to reconfigure cognitive categories and relational dynamics. In essence, then, border techniques are active around, across and within the sensing and computing environments, and constitute an extensive infrastructure of data sourcing, identification and profiling. These have been widely documented in literature, along with concerns on their potential political implications. These concerns have been often registered under concepts of surveillance and dataveillance (Kitchin 2014; Tufeckci 2014). Smart cities, David Lyon (2018) argues, bring along the normalization of surveillance, and metaphors like ‘the new panopticon’ (McMullan 2015) or the ‘big brother city’ (King 2016) have been mobilized in the media to describe cities governed from dashboards, where data about everyone and everything is gathered all the time and anonymity becomes impossible.
My intention is not to deny that cities are sites where dataveillance is particularly concentrated. I argue nonetheless that dataveillance is not an exhaustive framework for the analysis of data-driven urban governmentality, for two main reasons. First, despite the efforts of smart city planners, dataveillance often fails. The infinite amount of data gathered through sensing infrastructures does not automatically translates into government actions. Data are often dispersed among several different actors (states, municipalities, private firms, academic or non-academic researchers, NGOs, activists, hackers, etc. etc.) which pursue different and often conflicting agendas. This creates zones of opacity. Urban data can be so immense and fragmented that their potential in terms of actual, actionable knowledge remains largely under-exploited. Paradoxically, there might be so much dataveillance, that it makes complete dataveillance impossible. In short, data largely go wasted; or maybe, big data as such is waste, until it is dissected by algorithms, and reassembled in the form of actionable information. This is one of the problems that smart city projects like New Town are trying to address, by creating central control platforms.
But even if dataveillance is applied to the fullest extent, and no data is wasted, it still does not define a logic of urban government. Dataveillance accounts for some important aspects of data-driven environments; it is a disposition (Easterling 2010) of the sociotechnical assemblages we live in. But, as such, dataveillance does not explain how decisions are taken or strategies take form. Against the common emphasis on the big of big data, Louise Amoore and Volha Piotukh (2015) demand attention for the work of little analytics in contemporary forms of knowledge production and government. Through specific practices of data ingestion, partitioning and memory, the heterogeneity of life is flattened and reduced to patterns of data that are tractable for commercial or security decisions. This is exactly the logic of urban platforms like New Town. These work for urban security not by monitoring more, but by translating what is monitored into models, such as risk alerts, and possible actions. Paradoxically, data scientist and officers in the urban control rooms might be better off with less data, but sharper analytics, than with more data without an algorithmic way through. Dataveillance does not explain new forms of urban government because it keeps the focus on the aspect of watching and on the accumulation of data, while overlooking the specific operations – scraping, skinning, connecting, drawing and, ultimately, modelling – through which algorithms make data actionable and inform decisions.
This chapter has illustrated how smart city planners in New Town seek to forge a system of urban government where, not too differently from what happens at smart borders, algorithmic calculations launched across different sets of urban data provide city officers with profiles of the performances of citizens, transports, traffic, emergency services, weather, resources, pollution, and so on. The analytics chain elaborates these data to create models of future events. In the vision of smart government, these models are the grounds for political and administrative operations. Independently of governmental projects, the same activity of profiling and modelling is undertaken by private actors, such as IoT and software providers, for commercial purposes. My point here is that the border techniques ubiquitously incorporated in urban smart technologies form a pre-emptive apparatus. This is not limited to surveillance functions and frames a specific modality in which urban government is conceived and performed. Benedict Anderson (2010) identifies pre-emption as one of the logics of anticipatory action – together with precaution of preparedness – whose specificity is that it works on undetermined, potential scenarios of the future, and that increasingly defines government in our time. Pre-emptive governance seeks to incorporate not the probability, but the imagination of future possibilities, into security procedures (De Goede 2012; Amoore 2013). Security, then, has become speculative (De Goede et al. 2014); algorithms do not predict, but think through data and build models of the future upon which present action can be taken. From this perspective, borders built within sensing/computing technologies appear as the (sometimes involuntary) infrastructure of new strategies of urban government, whose effects are only becoming to unfold.