Smart cities and the imagined objectivity of data
Whether it is about asking your smart home device to close the shutters, selecting the right employee for a vacancy, detecting criminals, or reducing pollution: smart technologies are finding their way into the mainstream society. This article seeks to examine the following question: What is the influence of data aggregation and data-based urban policy on the residents of Smart Cities?
Why Smart cities?
One of the reasons for the growing popularity of smart technologies could be that they are driven by artificial intelligence (AI) and are therefore capable of making sense of big data sets (Kitchen, 2014). Those data sets can be derived from different systems such as surveillance cameras, heating or noise sensors, and applications such as Google Maps which makes visible how many people are at a certain place at a specific time. Another reason for the popularity of smart technologies might be that they are assumed to reduce, or maybe even end, arbitrariness and prejudice in knowledge construction and decision-making, in contrast to their human counterparts (Kitchin, 2014).
During the last twenty years there has been a strong tendency to move from rural areas of the world towards more urban places (Yu & Xu, 2018). According to Damori (2014), more than 52% of the world population lived in cities in 2014. As a result of the mixture of the trends of urbanization and smart technologies, Smart Cities have received more attention as well (Yu & Xu, 2018). Kitchin (2014) elaborates on the concept of Smart Cities and describes it as full of digital and technological devices that are built into its environment. He states that those built-in digital devices are used to:
“monitor, manage and regulate city flows and processes, often in real-time, and mobile computing (e.g., smart phones) used by many urban citizens to engage with and navigate the city which themselves produce data about their users (such as location and activity).” (Kitchin, 2014, p. 2)
All over the world countries are piloting the adoption of Smart Cities
The residents of Smart Cities are therefore being monitored on specific aspects of their daily lives. Without them actively sharing information, information about them is generated automatically and the agency therefore shifts from the individual to the one doing the monitoring. This collection of data, eventually, is used to gain a more in-depth understanding of the particular city and to design supposedly value free and evidence-based urban policies (Kitchin, 2014). This, however, might have consequences for the residents of the Smart City. This article seeks to examine the following question: What is the influence of this data aggregation and data-based urban policy on the residents of a Smart City?
In order to answer this question, the first section of this article focusses on several examples of smart cities and the institutions gathering the data. Thereafter, the consequences of the embedded technologies and new forms of surveillance will be discussed. The last section of the article will question the claimed objectiveness of data processing.
Same same but different
All over the world countries are piloting the adoption of Smart Cities, and in total there have been around thousand Smart City pilots. China, undoubtedly, takes the cake with a total of 269 ongoing pilots (Yu & Xu, 2018). In Hangzhou, for instance, there is a project called City Brain. City Brain was created by the multinational company Alibaba and it consists of different projects. One of them is Community and Public Safety, where they use video analysis technology and video recognition algorithms to take preventive measures to ensure the safety and security of the public (GOV Insider, 2018).
Another project within Alibaba’s City Brain is Traffic Congestion and Signal Control in which surveillance cameras and sensors are used at different places in the city to collect information on road conditions in real time (GOV Insider, 2018). Subsequently, this data collected in real time is sent to an AI Hub, in which traffic signals throughout the whole city are managed. Similar projects can be found in other Chinese cities such as Shanghai with their Citizen Cloud, the Huawei Pay App from Beijing, and the Smart Medical App from Guangzhou.
As can be seen from the examples above, every city has its own focus and the data collection can be carried out by both public and private institutions (Damori, 2014). This raises questions about ethical issues because who can be held responsible for the big data set if its sources consist of all kinds of different sensors enabling the data aggregation? The fact that every smart phenomenon is one of its own is also a complicating factor in Smart Cities, because this might mean that it is not evident for residents where they are being observed and what the collected data is used for. The consequences of these forms of data collection and monitoring for people in Smart Cities will be discussed in the next section.
Track and Trace
As might have become evident already, in Smart Cities people's actions and interactions are being tracked across a number of domains (Kitchin, 2014). Being in a public space, for instance driving a car on the highway or walking through the city in Hangzhou, exposes the individual to surveillance technologies while the government and/or private institutions are the ones looking. The residents, however, do not know exactly when someone is watching them because the technologies are embedded in the urban environment. As a consequence, residents are more likely to act in a socially accepted way in order not to be deviant or stand out (Brighenti, 2007).
The residents, however, do not know exactly when someone is watching them because the technologies are embedded in the urban environment.
In that sense, the Smart City can be compared to a panopticon (Kitchin, 2014). A panopticon is a building designed to allow all people in that building to be observed, without the person doing the observing being visible for the other people in the building. Therefore, the people in the panopticon will behave as though they are being watched because they never know when someone is observing them. Therefore, this new form of data aggregation might also have the implicit function of directing people into certain normative behavior.
This connects seamlessly with the concept of orders of visibility from Hanell and Salö (2017). Hanell and Salö (2017) state that social structures in society determine how knowledge is stratified and which knowledge will be visible and which knowledge will not. In that sense, knowledge and power cannot be seen separately from each other because the person with power determines which knowledge should be considered to be more legitimate, and knowledge which is more legitimate is perceived to have more power (Foucault in Hanell & Salö, 2017). In both examples, the Smart City and the panopticon, the person watching is the person with the power because they are also the one deciding which behavior is considered to be the norm.
It should be taken into account, however, that all knowledge is socially constructed and therefore dependent on many factors (Berger & Luckmann, 1966). This socially constructed nature of knowledge is often forgotten or excluded when talking about Smart Technologies. Take for instance Hangzhou’s City Brain project ‘Community and Public Safety’, in which camera surveillance and video recognition algorithms are used to determine where an unsafe situation is taking place in order to allow authorities to take preventive actions. In this Smart City project, it is not clearly explained what is considered as an unsafe situation.
Although the technologies are smart, the norms cannot be decided by the data or algorithm itself, as these decisions are made by humans first. Taken into account the socially constructed nature of knowledge, this means that data and/or algorithms are not as objective as they are presented to be. The next section will elaborate further on the concept of data objectivity.
Don’t blame us, it is the data!
People making use of smart technologies and algorithmic data often use the claim that their data is objective because there is no human inference in the data gathering (Gillespie, 2014). A camera or a sensor itself does indeed have no political agenda or prejudices. This makes projects like Smart Cities look like they are based on objective knowledge. However, an algorithm cannot function without human inference because there always need to be ideas, people and context to store them (Kitchin, 2014).
First, it needs to be decided which data should be gathered. The surveillance cameras used at the City Brain to track the traffic in Hangzhou are not asked to gather only data about the color of the clothes the people in the car are wearing; they are asked to gather data about for instance the speed of the cars and so on. Those decisions can only be made by humans in the first place. Additionally, the data needs to be made algorithm-ready in order for the algorithm to make sense of it. Moreover, the data gathered in real time from the surveillance cameras or other embedded smart technologies is used alongside data gathered over time (Kitchin, 2014).
The data control centers, such as the AI-Hub in Hangzhou, then integrate and bind together different sets of data to one big data set which is eventually used to make decisions upon (Kitchin, 2014). The decisions will therefore always depend on presuppositions regarding what matters for a specific goal. As Gillespie describes: “from this perspective, we might see algorithms not just as codes with consequences, but as the latest, socially constructed and institutionally managed mechanism for assuring public acumen: a new knowledge logic.” (2014, p. 192).
Algorithmically gathered data is never fully objective
The example of the color of the clothes of the people in the car that I just gave, might not sound very disturbing as an object of data collection. However, such decisions are also made with regards to more sensitive issues, such as Community and Public Safety in which it needs to be decided what is considered to be an unsafe situation. In that specific case it might be possible that specific groups in society are expected to be more threatening to public safety. This is then again based on socially constructed knowledge and subsequently the video recognition algorithms are taught to look out for those certain groups even more. This, eventually, could lead to disadvantages for specific groups in society because they are made visible in a specific way that is legitimized by the powerful ones (Hanell & Salö, 2017).
In summary, algorithmically gathered data is never fully objective. However, this claim of objectivity is often used by developers and policy makers in order to guarantee that societal decisions are based on what appear to be common-sense and evidence-based findings.
The data-based urban policies of a Smart City
I set out to determine the influence of data aggregation and data-based urban policies on the residents of a Smart City. What we have seen is that the data gathered in Smart Cities is not only used to optimize people's lives and to stimulate economic development, but that it could also lead to politics of inclusion in which it is determined how an individual should behave in relation to the population (s)he is part of. People then modify their behavior because it is not always evident for them where they are being observed and what the collected data is used for.
Another factor that influences the residents of the Smart City is that data gathered in Smart Cities is not actually as objective and innocent as it looks like, because this data cannot exist without human interference. The institutions gathering the data, which are both public (governments) and private (i.e. Alibaba, Huawei), have the power to decide which information is used to build urban policies on (Schuilenburg & Peeters, 2018).
This imagined objectivity of the data and ignorance regarding these new forms of knowledge in the current age, gives governments and the people deploying smart technologies the opportunity to describe their findings as objective and accountable. This might let people just take the knowledge as it is and forget about the socially constructed nature of it. Perhaps we can even say that in this form, people are served a mirage, but in fact this can lead to new forms of surveillance and directing people towards conforming to certain normative behaviors.
Brighenti, A. (2007). Visibility. Current Sociology, 55(3), 323–342.
Dameri, R. P. (2014). Comparing Smart and Digital City: Initiatives and Strategies in Amsterdam and Genoa. Are They Digital and/or Smart? Smart City, 45–88.
Gillespie, T. (2014). The Relevance of Algorithms. In T. Gillespie, P. Boczkowski, & K. Foot (Eds.), Media Technologies: Essays on Communication, Materiality, and Society (pp. 167–194). Cambridge, United States: MIT Press.
GOVInsider. (2018, October 17). Five Chinese smart cities leading the way | GovInsider. Retrieved December 5, 2018.
Hanell, L., & Salö, L. (2017). Nine Months of Entextualizations. In C. Kerfoot, & K. Hyltenstam (Eds.), Entangled Discourses (pp. 13–61). New York, United States: Taylor & Francis Group.
Kitchin, R. (2014). The real-time city? Big data and smart urbanism. GeoJournal, 79(1), 1–14.
Kitchin, R. (2015). Making sense of smart cities: addressing present shortcomings. Cambridge Journal of Regions, Economy and Society, 8(1), 131–136.
Luckmann, T., & Berger, P. (1966). The foundations of knowledge in everyday life. In T. Luckmann, & P. Berger (Eds.), The Social Construction of Reality (pp. 13–61). New York, United States: Penguin Random House LLC.
Schuilenburg, M., & Peeters, R. (2018). Smart cities and the architecture of security: pastoral power and the scripted design of public space. City, Territory and Architecture, 5(13), 1–9.
Yu, W., & Xu, C. (2018). Developing Smart Cities in China. International Journal of Public Administration in the Digital Age, 5(3), 76–91.