Towards a New AI Urban Regime

Towards a New AI Urban Regime

The expansive AI regime, within the smart city paradigm, represents the maturation of both the information economy (where maximizing data and information is a key driver of economic value) and the expansion of the so-called "Fourth Industrial Revolution" (where autonomous decision-making by machines promises to improve efficiency and effectiveness in all types of production processes). The evolution of "smart cities" towards a widespread use of AI techniques, processes and devices indicates that a new urban AI regime will expand in the coming years, and therefore it is necessary to make an analysis of both the benefits and the risks of this strategy.

We can define an urban AI regime as a stage of evolution in urban growth machines, characterized by a set of formal and informal agreements between public agencies and private interests that (1) prioritize development of smart cities, particularly algorithm-based technologies, and (2) present the necessary institutional frameworks that allow the realization of the goals and objectives of the AI industry, which would supposedly benefit all segments of the urban population.

AI is a collection of algorithms programmed to mimic human decision making. It refers to a variety of computer systems and applications that can “perceive” their environment, think, learn, and act in response to what they perceive and their programmed goals. AI is relatively useless without a set of intentional goals to complement it. Therefore, the initial task facing people who plan, build, and manage AI systems is to determine the kind of outcomes they want machine learning algorithms to pursue (Tomer, 2019).

Broadly speaking, four key components can be identified in smart city processes: data sets, communication platforms, timely decision-making, and effective action. As a specific technology, AI adds to smart city processes some key features: machine learning abilities, automation processes and autonomous decision-making, and, with so-called deep learning, forecasting capabilities that can be combined with digital twinning and other digital simulation techniques (Berryhill et al, 2019).

We researchers recognize that artificial intelligence (AI) represents one of the fundamental elements in the development of smart cities. Along with machine learning, AI is well suited to form the analytical foundation of smart city programs. Research in this field is still in its infancy. Artificial intelligence is expected to become a core element of an increasing proportion of smart city applications.

AI is already playing a role in contemporary high-tech clusters and spaces, mainly in inno-districts and factories of the future, but increasingly in science cities, smart ports, production service complexes, industrial eco-parks, etc These spaces share with traditional technopoles a concentration on innovative and advanced manufacturing, as well as information-based industries. Some inno-districts combine AI and robotic platforms with ecosystems of start-ups and hi-tech companies relocating to city centers in the last 10-15 years.

Research and innovation in advanced manufacturing taking place in the “factories of the future”, located in urban areas and metropolitan regions, is making extensive use of devices and autonomous decision-making processes to optimize production. An example is the so-called lights out manufacturing, where production sites are fully automated and processes such as CNC (computer numerical control) machining do not require human presence.

Some specific urban areas in numerous cities are becoming AI production spaces where programming and algorithm meetups (hackathons) are organized to produce code. Some inno-districts host these meetings, which are organized as design thinking events. The goal of a hackathon is to create functional software or hardware during the event. According to critics, sponsors take advantage of the free labor of young engineers and programmers to create “fictitious expectations of innovation” (Griffith, 2018).

Destructive Creation

Despite the incremental nature of smart city implementation, its disruptive character is undeniable. Undoubtedly, disruptive innovation takes place within the general framework of capitalist creative destruction, as Schumpeter showed us. This disruptive impulse adds to the development process of urban megaprojects as an essentially disruptive and contentious enterprise, as we have shown in some specialized publications (del Cerro Santamaría, 2019). 

The disruptive urban regime of AI (which we describe as destructive creation) generates a critical mass of negative externalities in the development process of smart cities, derived initially from the contemporary hegemonic nature of technological innovation in urban socio-economic processes. One consequence of this is the lack of exploration of alternative trajectories for urban development based on other models, values, priorities and regulatory frameworks.

This scientifically planned strategy, which breaks down all alternatives and makes AI a necessary and unavoidable component of urban development (potentially leading to monopolization), is a major threat to the chaotic and random character of cities, as Richard Sennett has always argued (Sennett, 2012). It is also a risk for the centrality of public spaces and the civic friction between humans that should occur, according to Jane Jacobs, in the core of the urban as self-organized complexity (Jacobs, 1961).

On the other hand, dependence on a single, comprehensive and interrelated set of strategies for urban development, offered by a single industry, the technology industry, presents serious risks derived from the mobility of capital. Since the business location process is exclusively decided by technology corporations, private interests can determine the success and failure of urban policy and urban development. The failed negotiations in New York City in 2019 to try to get Amazon to build a new corporate headquarters in the borough of Queens is a case in point.

Big tech, in fact, needs to be scrutinized. The production mode promoted by Facebook, Google and Amazon claims to work with "artificial intelligence". In reality, however, it is held together by tens of millions of anonymous workers in warehouses, data centers, factories, electronic assembly shops, factory farms, and processing plants where they are left unprotected from disease and hyper-exploitation (Klein, 2020).

As a top-down corporate strategy, the nascent AI regime will likely foster a race for inflated valuations in processes, spaces, devices, and outcomes that would tip urban economies toward a greater focus on networked land and intellectual property rights.

Technological gentrification (interrelated with green gentrification) is already an emerging feature in spaces adjacent to innodistricts and other urban tech neighborhoods. As a consequence, large segments of the urban population (those living in poverty, without access to basic services or with accessibility problems due to disabilities) can potentially be left behind (Watson, 2013).

These are some of the shortcomings of the political economy of the nascent urban regime of AI. Other risks and ethical dilemmas brought about by the smart city strategy have to do with control, privacy and security. This occurs in a “surveillance” situation stemming from the high level of big data collection, analysis, scanning, and identification for predictive monitoring, as Zuboff points out in her book on Surveillance Capitalism. We know that there are also cognitive risks, such as “intellectual automation” and “automation complacency”, according to the research by Nicholas Carr (Carr, 2014).

On the other hand, “deep learning” operates on a black box model that is not always understandable to humans. AI could potentially be hacked, allowing cybercriminals to interfere with energy, transportation, early warning, or other crucial systems. Since AI systems interact autonomously, they can produce unpredictable results, and thus uncontrolled AI could pose an existential danger to humans . Last but not least, AI-driven automation can lead to massive job losses in any industry (Muro et al, 2019).

AI Innovation in Context

Improving data efficiency by optimizing big data management is not, in and of itself, a substitute for good policies. The pursuit of urban sustainability would benefit from (1) a comprehensive policy of sustainability in the growth machine that encompasses the various components and objectives of sustainability and (2) an ex-ante collective deliberation of the results that will be achieved through the implementation of AI strategies.

The lack of a normative approach and the controversies around data ownership are two characteristics of smart city processes. It is not clear how the recent evolution towards the incorporation of artificial intelligence technologies in smart cities can contribute to mitigate the deficiencies and the controversies that have been generated.

Ultimately, the values and goals of human development should determine what technological innovations we need. Put differently, technological innovation and AI by themselves do not necessarily foster urban sustainability, even though some specific innovations (technological or otherwise) may. Furthermore, in the case of those innovations and AI technologies that promote sustainability, it is necessary to take into account both their benefits and their risks and, therefore, seriously consider issues of resilience or “anti-fragility”, as Taleb (2012) points out. . 

The key question, therefore, is how to design and build a framework for sustainability that is capable of discriminating and showing which AI innovations can contribute to the sustainability outcomes we seek, rather than starting the path to sustainability from the innovaton side. As in the case of the idea of “quality of life”, the concept of sustainability (along with the United Nations sustainable development goals) probably needs to be redefined, articulated and specified. In this process of conceptual and practical clarification, both common global trends and specific local conditions play an important role. The imagery of sustainability is that of a complex set of assemblages. This suggests multiple interrelated components and scales of action in a baroque and rhizomatic structure that invites analysis through both complexity and transdisciplinarity (del Cerro Santamaría, 2020).

Instead of settling for the kind of sustainability that potentially results from AI urban innovations, we need to review the design of the sustainable urban future we want, and then lay out its main components and requirements. In order to promote urban sustainability we need to assess the goals that AI decision tools are trying to accomplish. In fact, the link between artificial intelligence and sustainability would require an explicit policy and regulatory framework. Such a framework would specify priorities and expected outcomes so that we can understand precisely under what conditions AI innovation leads to sustainable urban scenarios and outcomes.

The nascent AI regime should also be examined alongside fundamental questions about (1) what benefits megaprojects bring to cities, (2) which projects should be built and which should be scrapped, and (3) how to mitigate the broader socioeconomic impacts of large-scale development projects. To be sure, AI and innovative technologies would contribute more effectively to improving the quality of management or the sustainability of projects if they were framed within a solid process of analytical clarification and construction of grounded theory with respect to priorities and objectives that are based on evidence (Köhler et al, 2019). 

Tested results on security, explainability, transparency and validity would contribute to improving the implementation processes of artificial intelligence and smart cities in the planning and construction of urban projects. In addition to concerns about improving project management and urban development through high-tech innovation and AI, researchers need to develop analytic tools to account for urban projects as socioeconomic particles producing hugely negative externalities and social impacts. 

Intelligent disruption (destructive creation) produced by Artificial Intelligence systems is no longer avoidable. We must try to tackle it via relevant debates around what kind of sustainability we want for ourselves and for future generations. These should be discussions around the notion of smart and fair sustainability, a normative horizon from which we could decide what technological innovations and what AI applications we need.

References

Berryhill, J., Heang, K., Clogher, R., K. McBride (2019) Hello, World: AI and Its Use in the Public Sector, OECD Working Papers on Public Governance, no 36, https://dx.doi.org/10.1787/726fd39d-en.

Carr, N. (2014) The Glass Cage. Where Automation Is Taking Us, New York: W. W. Norton.

del Cerro Santamaría, G. (2019) Disruptive and Contentious Enterprises: Megaprojects in Bilbao, Istanbul and Hong Kong. In XIII CTV 2019 Proceedings: XIII International Conference on Virtual City and Territory: “Challenges and paradigms of the contemporary city”: UPC, Barcelona, October 2-4.

del Cerro Santamaria, G. (2020) Innovation Districts, Complex Sustainability and Urban Redevelopment. Evaluating Multiple Success Factors, IJEBMR 4 (6), 279-290.

Griffith, E. (2018) Sociologists Examine Hackathons and See Exploitation, Wired (Business Section), March, https://www.wired.com/story/sociologists-examine-hackathons-and-see-exploitation/. Accessed December 2021.

Jacobs, J. (1961) The Death And Life Of Great American Cities, New York: Random House.

Klein, N. (2020) Screen New Deal,The Intercept.

Köhler, J., F. Geels, F. Kern, J. Markard (2019) An agenda for sustainability transitions research: State of the art and future directions, Environmental Innovation and Societal Transitions 31. Accessed October, 2021.

Muro, M., R. Maxim, J. Whiton (2019) Automation and Artificial Intelligence. How Machines Are Affecting People and Places, Brooking Institution Report, January, https://www.brookings.edu/wp-content/uploads/2019/01/2019.01_BrookingsMetro_Automation-AI_Report_Muro-Maxim-Whiton-FINAL-version.pdf. Accessed November 2021.

Sennett, R. (2012) "No one likes a city that's too smart," The Guardian. Archived from the original on 18 March 2017. Accessed 17 March 2017.

Taleb, n. (2012) Antifragile. Things That Gain from Disorder, New York: Random House.

Tomer, A. (2019) Artificial intelligence in America’s digital city, Brookings Institution report, https://www.brookings.edu/research/artificial-intelligence-in-americas-digital-city/, Accessed September 2021.

Watson, V. (2013) African urban fantasies: dreams or nightmares?, Environment and Urbanization. 26 (1): 215–231.

Author

Gerardo del Cerro Santamaria

Gerardo del Cerro Santamaría

New York City, United States of America

Professor of Planning and Megaprojects at The Cooper Union in Manhattan for 22 years. Invited Professor at MIT, Columbia University, London School of Economics and University College London. Director of Evaluation at the U.S. NSF (Gateway Engineering Program). U.S. Fulbright Award Recipient. Author of over 90 academic publications. His new book is Megaprojects in The World Economy. Complexity, Disruption and Sustainable Development (New York: Columbia University Press, 2023).

Photo: Cloud Valley, an AI-based city in southeast China, is planned to occupy more than a million square meters

 

 

We use our own and third-party cookies to enable and improve your browsing experience on our website. If you go on surfing, we will consider you accepting its use.