"One AI to rule them all": The Unification of Chinese Urban Governance under

The premise: artificial intelligence and the unification of urban governance

We live in a time when the development of cities cannot be understood anymore as the sum of different and disconnected processes carried out by diverse intelligences. Urban development has always been an extremely complex and multifaceted activity comprising, for example, the creation of new urban spaces and the maintenance of the existing built environment, the preservation of the natural environment, the attraction and investment of capital and the fulfilment of citizens’ social needs. Because of the sheer diversity of spaces, activities and services that underpin urban development, traditionally its governance has been operated by different actors in charge of different urban domains. This of course does not mean that urban actors have performed their job in silos completely disconnected from central governments and from other urban actors. Yet there is a fundamental difference, for instance, between an Environment Bureau in charge of protecting local ecosystems and a Public Safety Department meant to ensure the protection of citizens. The difference lies in the fact that, in this example, different urban services are delivered by different human actors who have an impact on different urban spaces. In this paper, I argue that this condition is being radically altered by artificial intelligence (AI). In the following sections I discuss the case of China’s urbanization to illustrate how, what was once a multiform urban governance characterized by a plethora of different actors, offices and spaces, is now converging towards a single artificial intelligence and a single cyberspace. We begin our exploration with China’s environmental policy and ecological ambitions, to see how the environmental governance of Chinese cities is going beyond physical spaces and human actors.

The case: the platformization of China’s urbanization

In 2020, the Chinese government announced an ambitious environmental programme aiming to peak carbon emissions by 2030 and to reach carbon neutrality by 2060 ¹ . This broad national programme reflects a number of environmental initiatives that are taking place in cities which are both the engine of China’s economic growth and the cause of its environmental problems ² . For example, recent studies show that while Chinese cities are responsible for 75% of the national GDP, they are also responsible for more than 70% of the whole country’s carbon emissions, and for significant loss of natural habitat. It is in this complex urban context, in which economic priorities clash against environmental dilemmas, that AI is emerging in China as an instrument to manage the country’s rapid urban expansion and to decarbonize its cities. Innovation in AI was pushed forward in China by the State Council (China’s chief administrative authority) in 2017 with the publication of the New Generation of Artificial Intelligence Development Plan, and it is now culminating in the creation of so-called city brains ³ . A city brain is a type of urban artificial intelligence meaning an AI that is capable of acting autonomously in urban environments, on the basis of information acquired through sensory experience. More specifically, city brains are urban AIs located in digital platforms through which they manage large portions of urban governance, ranging from transport to safety and from environmental preservation to urban planning.

Although similar urban AIs exist in other parts of the world, city brains are a Chinese invention. The first city brain was developed in 2016 by the Chinese tech giant Alibaba which installed it and tested it in the city of Hangzhou in the Zhejiang province. Alibaba’s city brain is emblematically called City Brain and its functions and applications exemplify the characteristics of this urban AI and its impact on urban governance. City Brain’s original function pertained to the management of Hangzhou’s traffic. This urban AI is capable of sensing the surrounding environment by means of hundreds of CCTV cameras distributed in the city ⁴ . These cameras act as the eyes of the city brain: they observe what is happening in the city, by producing data in real-time. This real-time data is combined, in the city brain’s digital platform, with big data in the shape of, for example, urban maps and weather forecasts. Once blended, this vast pool of data allows City Brain to develop a situational awareness regarding the traffic conditions of Hangzhou ⁵ , for policy makers to intervene in the present and change the future mobility of the city. In practice, City Brain can predict when and where traffic congestion is likely to occur, and it autonomously optimizes the city’s traffic lights to avoid the formation of traffic jams ⁶ .

The role of City Brain in the governance of urban transport and mobility resonates with China’s environmental ambitions. According to Alibaba, the application of this urban AI has already decreased traffic congestion by 15% and reduced travel time by 8% ⁷ . Provided that car ownership does not increase substantially, these results can correlate positively with reductions in carbon emissions, since the new mobility patterns established by City Brain effectively means that, in a city like Hangzhou, cars get stuck in traffic for less time and are able to complete their journey faster. This in turn means that, as cars are in motion for less time, overall they generate less carbon emissions which is exactly what Chinese policy makers need to achieve as soon as possible, in order to decarbonize the whole country by 2060.

This is only the first part of the story. As mentioned earlier, city brains reside in digital platforms which, as the literature on platform urbanism shows, are being increasingly used all around the world to manage urban services ⁸ . Digital platforms have a peculiar architecture that makes them highly versatile and compatible with numerous types of data and services ⁹ . Essentially a digital platform has a modular structure allowing for easy assembly and flexible arrangement, beyond the original purpose and capabilities. In the case of City Brain, this means that the digital platform where this urban AI operates can be easily expanded by Alibaba to incorporate functions that go beyond the management of traffic. Once again there is a connection with China’s environmental programme, since part of the evolution of City Brain has been perfectly in line with the country’s aim of decreasing carbon emissions and preserving the natural environment. For example, a new component of City Brain’s digital platform is called Environment Brain which combines a geolocation system with environmental data to anticipate how much waste will be produced, predict the capacity of photovoltaics and foresee the carbon footprints of private companies ¹⁰ . The logic is the same as the one that was originally applied in Hangzhou to optimize traffic. The city brain collects data and employs it to develop predictions about what is likely to happen in a given urban system, for that system to be optimized in advance. The technology is also similar. Sensors are employed to develop a situational awareness of the city, as it is in the present, and then the AI predicts its future. What differs is the aspect of the city that is taken into account in the analysis of the present and the anticipation of urban futures. Originally, the urban aspect that City Brain was focusing on was traffic. Later, the AI’s focus was extended to environmental aspects such as waste, clean energy and carbon footprint, by means of the flexible architecture of the digital platform that allows Alibaba to plug in new analytical and predictive features related to more and more urban domains. But where does the expansion of City Brain’s digital platform end? And where does the agency of this urban AI stop?

The answer to both questions is: it does not stop. The modular structure of a digital platform is purposely designed to grow ad infinitum. New components, features and capabilities can, from a technical perspective, always be added, provided of course that they are compatible with the original logic and technology. On these terms, Alibaba can keep expanding City Brain, provided that its new functions relate to the accumulation of data about present cities and the algorithmic predictions of urban futures. Moreover, it is important to remember that, although a city brain has physical components such as computers and sensors, its core essence is digital and located in an infinite cyberspace where endless growth is theoretically possible. City Brain’s digital platform can thus keep growing without any immediate physical limit and so its agency since, for every new feature and component that is plugged in, this urban AI extends its agency to another aspect of the city.

This is exactly what has been happening with Alibaba’s City Brain. This city brain originally built to manage urban traffic, was then expanded to manage domains of environmental governance, such as waste and energy, and it has now been further expanded to control aspects of urban governance, that include healthcare, supply chains and finance. Furthermore, the same AI that was once operational only in Hangzhou, can now be found in the governance of over twenty cities in and beyond China, including Kuala Lumpur in Malaysia. The growth of this city brain and that of its agency in urban governance do not stop.

The repercussion: one AI to rule them all

The platformization of China’s urbanization has just begun. What this terminology seeks to capture is the increasing employment of digital platforms in the governance of the many aspects that underpin the development and life of cities, from their waste to their mobility and from their carbon emissions to their finance, health and security. The list of these urban aspects is long and it is being quickly covered by the agency of city brains which are now emerging as leading actors in the governance of Chinese cities. This trend can be observed by looking at the specific case of Alibaba’s City Brain, discussed above, but it is also observable in three broader trends that characterize contemporary China. The first one is the diffusion of digital platforms which are being increasingly used in the Chinese society to manage a wide and diverse array of data ¹¹ . The second one is the practice of smart urbanism almost all over urban China, which with its emphasis on sensors and data to optimize urban governance, represents the urbanistic antecedent of platform urbanism ¹² . The third one is the push of the Chinese government for research and development in the field of AI, with the aim of establishing China as an AI superpower ¹³ .

From an environmental perspective, AI-enabled urban platforms are a double-edged sword around which there is a growing academic debate ¹⁴ . On the one hand, some scholars argue that environmental monitoring and waste management (which are among city brains’ capabilities) can be enhanced by AI ¹⁵ . Similarly, with road traffic accounting for a substantial portion of global carbon emissions and for approximately 10% of China’s CO2 emissions, its autonomous management and optimization via a technology such as City Brain cannot fully decarbonize cities, but it can indeed contribute to that goal ¹⁶ . On the other hand, it is unlikely that AI will fix urban issues of biodiversity loss, simply because in China, for instance, these problems are caused by the expansion of urban spaces that end up consuming and replacing natural spaces. AI is not a magic wand capable of recreating nature out of thin air. In addition, it is crucial to remember that AI consumes nature in the first place. Producing AI technologies requires mining critical minerals and metals in and outside China ¹⁷ . Most importantly, as this article has discussed, environmental features are just one module of digital platforms’ modular structure. Within these digital systems, no module operates in isolation. Interconnectivity is key to their mechanics and, with the platformization of the urbanization of China set to continue apace, there are three broader interconnected repercussions that we need to consider.

First, there is the issue of data blending. Chinese city brains harvest large amounts of data from a variety of sources. Images of people and places captured by CCTV cameras. Levels of energy production and consumption calculated by smart meters. Patterns of urban mobility understood via tracking systems. Carbon emissions measured by smart sensors. But also private information that was recorded in the past in government databases, as well as real-time and potentially sensitive information that almost every Chinese citizen is now producing by means of social media and mobile apps. From different corners of China and from different parts of its cities and society, all this data eventually ends up in the same digital platform. Within this platform, the origin of each piece of data and the original purpose for which it was produced and collected, do not matter anymore. Every bit of data gets blended and can be repurposed for aims that go well beyond the initial urban domain from which the information was coming from. This of course is an issue that is not exclusively Chinese. Quite the opposite, in fact. Under the influence of powerful digital companies based in the West, such as Google and Facebook, societies from all over the world are seeing their personal data be repurposed to foster targeted advertising, predictive policing and, in a word, surveillance ¹⁸ .

Second, there is the issue of the understandability of urban AI and of the digital system in which it operates. The platformization of China’s urbanization means, in practice, that a lot of data and services are moving to the digital realm. Almost paradoxically, cyberspace is nowadays immediately accessible from almost every urban area, including our domestic spaces, through a simple smartphone. Yet, cyberspaces are far away from physical spaces. They are located in different dimensions and their very fabric presents strong differences. These differences, together with the fact that a digital platform, such as the one where City Brain resides, is managed by a non-biological intelligence, make digital spaces and artificial intelligences very hard to understand by the majority of the population. This is an issue that has been perfectly captured in the literature on so-called XAI which stands for Explainable Artificial Intelligence ¹⁹ . In essence, what this strand of literature shows is that AI is an arcane and esoteric technology. The complexity of algorithms, how they are written and how they actually function, is such that those who do not have a background in computer science would struggle to understand their operation. Similarly, the modular structure of digital platforms and the seas of data that they ingest can be so complex to the point of being fully comprehensible only to data analysts. This is an issue because AIs and digital platforms are now being used to govern cities, thereby impacting real spaces that are part of our everyday life and, above all, our lives. There is thus a problematic asymmetry of knowledge in the condition of citizens who do not comprehend the AI that is shaping their life, while the AI itself, thanks to the immense blended datasets at its disposal, knows everything about them.

Third, there is the issue of centralization, intended as the concentration of power into a centralized AI. A city brain is a centralized AI in the sense that the power to know multiple aspects of cities, their lives and the life of citizens, as well as the power to actively govern cities and influence their development lies within a single AI, as opposed to it being distributed among various human and artificial intelligences. Of course, human agents continue to play an important role in the governance of cities. In China, for example, City Brain itself could not function without humans. It needs Alibaba’s computer scientists to build its digital platform and to write the algorithms that allow it to act. It also needs Alibaba’s data analysts to feed it with the data that is eventually used to develop predictions.And, last but not least, it needs human labor to build all the physical infrastructure, from computers to smart grids and from sophisticated sensors to simple cables, without which its digital realm could not exist. However, as the literature on urban artificial intelligence shows, cities are experiencing a passage towards autonomy. This is happening in and beyond China ²⁰ . Urban technologies and services that were traditionally operated by human agents are increasingly functioning in an autonomous manner. We are entering the age of the autonomous city: a city controlled by AI, in which more and more humans are left out of the loop. When the AI in charge is a city brain, the majority of power gets concentrated into one AI ruling not just over one city, but over many cities. Whether such omniscient, and potentially omnipotent, AI will turn into an enlightened absolutist or into a despotic tyrant is a question that needs to be urgently asked and answered before it is too late.

1. S. Mallapaty, “How China could be carbon neutral by mid-century”, Nature, 586, 2020.
2. K. Tang, Y. Liu, D. Zhou, Y. Qiu, “Urban carbon emission intensity under emission trading system in a developing economy: evidence from 273 Chinese cities”, Environmental Science and Pollution Research, 28(5), 2021 ; Y. Wang, X. Fang, S. Yin, & W. Chen, “Low-carbon development quality of cities in China: Evaluation and obstacle analysis”, Sustainable Cities and Society, 64, 2021 ; Y. Tu, B. Chen, L. Yu, Q. Xin, P. Gong, & B. Xu, “How does urban expansion interact with cropland loss? A comparison of 14 Chinese cities from 1980 to 2015”. Landscape Ecology, 36(1), 2021.
3. F. Wu, C. Lu, M. Zhu, H. Chen, J. Zhu, K. Yu, & Y. Pan, “Towards a new generation of artificial intelligence in China”, Nature Machine Intelligence, 2020.
4. F. Cugurullo, “Urban artificial intelligence: From automation to autonomy in the smart city”. Front. Sustain. Cities, 2(38) 2020.
5. F. Caprotti, D. Liu “Platform urbanism and the Chinese smart city: the co-production and territorialisation of Hangzhou City Brain”. GeoJournal, 2020.
6. J. Zhang, X. Hua, J. Huang, X. Shen, J. Chen, Q. Zhou & Y. Zhao, “City brain: practice of large-scale artificial intelligence in the real world”, IET Smart Cities, 2020.
7. Alibaba, “City Brain now in 23 cities in Asia”, 2020.
8. S. Barns, “Platform urbanism: negotiating platform ecosystems in connected cities”, Springer, 2019.
9. A. Lee, A. Mackenzie, G. Smith, P. Box, “Mapping platform urbanism: Charting the nuance of the platform pivot”, Urban Planning, 2020.
10. Alibaba, “ Environment Brain ”, 2021.
11. F. Caprotti, D. Liu, “Emerging platform urbanism in China: Reconfigurations of data, citizenship and materialities”, Technological Forecasting and Social Change, 2020.
12. A. Karvonen, F. Cugurullo, F. Caprotti (Eds.), Inside smart cities: Place, politics and urban innovation, Routledge, 2018.
13. H. Roberts, J. Cowls, J. Morley, M Taddeo, V. Wang, L. Floridi, “The Chinese approach to artificial intelligence: an analysis of policy, ethics, and regulation”, AI & SOCIETY, 2021 ; K. F. Lee, AI superpowers: China, Silicon Valley, and the new world order, Houghton Mifflin Harcourt, 2018.
14. T. Yigitcanlar, F. Cugurullo, “The sustainability of artificial intelligence: An urbanistic viewpoint from the lens of smart and sustainable cities”, Sustainability, 12(20), 2020.
15. C. Scoville, M. Chapman, R. Amironesei, C. Boettiger, “Algorithmic conservation in a changing climate. Current Opinion in Environmental Sustainability”, 51, 2021.
16. L. Zhang, R. Long, H. Chen, J. Geng, “A review of China’s road traffic carbon emissions”, Journal of Cleaner Production, 207, 2019.
17. L. Zhou, H. Fan, T. Ulrich, “Editorial for Special Issue ‘Critical Metals in Hydrothermal Ores: Resources, Recovery, and Challenges’ ”, Minerals, 2021.
18. S. Zuboff, The age of surveillance capitalism: The fight for a human future at the new frontier of power, Profile Books, 2019.
19. A. B. Arrieta et al., “ Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI ”, Information Fusion, 58, 2020.
20. F. Cugurullo, Frankenstein Urbanism: Eco, Smart and Autonomous Cities, Artificial Intelligence and the End of the City, Routledge, 2021.