Toward a ReThinking and ReImagining of Urban Sustainability in an Era of AI

Abstract

This paper provides a comprehensive review and analysis of the recent research and practice literature pertaining to urban and community sustainability while highlighting issues and potentials related to artificial intelligence (AI)-rich environments. The methodology for this paper includes a literature review drawing on a variety of sources (e.g., IEEE Xplore, Semantic Scholar, SpringerLink, etc.), a comparison of key focus areas with examples of challenges identified for each, followed by further analysis to focus on gaps and emerging initiatives. Key takeaways of this work point to a movement toward a rethinking and reimagining of urban and community sustainability in an era of AI; the importance of being aware of ever-emergent AI developments; and developing literacies that are crucial to navigating urban and community sustainability in an era of AI. In terms of concrete results, this paper provides a set of recommendations focusing on impacts, literacies, and policies.

1. Introduction

Contributing to the motivation for this paper is the claim made by Yigitcanlar and Cugurullo [1] that “[t]he merging of artificial and human intelligences in cities” gives rise to “the world’s next big sustainability challenge.” Zhang, Pee, Pan, and Liu [2] speak in terms of the “orchestrating” of artificial intelligence (AI) in support of urban sustainability in relation to resources as a theoretical lens while highlighting the importance of implementation. Concerned with the sustainable development goal (SDG) advanced by the United Nations (UN) of “fostering sustainable cities and communities,” Leal Filho, Mbah, Dinis, Viero Trevisan, de Lange, Mishra, Rebelatto, Ben Hassen, and Aina [3] explore the role of AI, pointing to gaps, obstacles, and potentials while also noting the importance of implementation. Lakusić, Pavičić, Stojčić, Scholten, and van Sinderen Law [4] explore the urban landscapes of Europe in a post-industrial age, providing an overview of challenges and opportunities associated with deindustrialization and the transformation of “the socio-economic, cultural, and physical fabric” of cityscapes aided by technologies such as AI and by “effective governance” in support of development that is sustainable. Weng, Li, Cao, Lu, Gamba, Zhu, Xu, Zhang, Qin, Yang, Ma, Huang, Yin, Zheng, Zhou, and Asner [5] explore the transformation potential of AI in relation to integration with earth observation (EO) and “urban observing, sensing, imaging, and mapping” finding “a deeper interpretation and autonomous identification of urban issues” enabling “the creation of customized urban designs.” Owojori and Erasmus [6] advance the notion of a metaverse platform (said to be transformative) for urban sustainability reporting, focusing on transparency, accountability, and communication “where users can interact with a digital representation of the real world.” However, Owojori and Erasmus [6] acknowledge the many challenges associated with the Metaverse platform, from cost to equitable access to data privacy, as well as ethical governance, to name a few. And it is perhaps worth noting that currently a call exists by an international journal on urban sustainable development for an article collection on “AI for smarter and sustainable cities” (IJUSD) [7] where planning, design, and implementation figure strongly.

As such, the purpose of this paper is to conduct an exploration of the research and practice literature in consideration of the following research question:

RQ. 

How is sustainability manifesting, particularly in an era of AI, in urban communities and regions and what are some of the associated challenges and opportunities?

To this end, this paper provides a review of the current state of the research and practice literature, highlighting gaps and issues associated with sustainable cities and communities, in an era of AI. This review paper is important in that it surfaces key areas of focus as well as challenges emerging in the research literature such as the importance of governance, integration, and implementation considerations for AI in relation to sustainable cities and communities. Key conclusions highlight the rich array of ways in which a rethinking and reimagining of sustainability in urban and community environments is occurring in an era of AI.

Definitions are provided in Section 1.1 for key terms used in this paper, based on a glimpse of examples and perspectives from the research and practice literature. Section 1.2 provides an overview of the objectives of this work and Section 1.3 provides an overview of the methodology guiding the exploration in this review paper.

1.1. Definitions

Agentic Urban AI. Tiwari [8] defines agentic urban AI as “urban-embedded intelligences capable of autonomously defining, prioritizing, and pursuing objectives that may diverge from their initial programming” consisting of the “defining features” of independent goal formation, strategic adaptation, behavioral cooperation, and value re-prioritization as well as contextual responsiveness, long term planning, minimal human oversight, and normative reasoning.

Artificial Intelligence (AI). Cugurullo, Caprotti, Cook, Karvonen, McGuirk, and Marvin [9] point to the many and varied forms of AIs, which share a range of common traits such as the artificial, learning capabilities, concept extraction, acting autonomously, and managing uncertainty, to name a few.

Generative Artificial Intelligence (GenAI). According to Clark, Barton, Albarqouni, Byambasuren, Jowsey, Keogh, Liang, Moro, O’Neill, and Jones [10], GenAI “includes a wide spectrum of artificial intelligence (AI) dedicated to creating or generating content or data that frequently resembles human-generated content.”

Smart and Sustainable City. While many definitions have been advanced for the smart cities concept (Nam and Pardo; Moura and de Abreu e Silva) [11,12], Yigitcanlar and Cugurullo [1], citing the work of Yigitcanlar, Hoon, Kamruzzaman, Ioppolo, and Sabatini-Marques [13] define a smart and sustainable city, for example, as “an urban locality functioning as a robust system of systems with sustainable practices, supported by community technology, and policy, to generate desired outcomes and futures for all humans and non-humans.”

Sustainability. The United Nations (UN) [14] defines sustainability, citing the work of the UN Brundtland Commission, as “meeting the needs of the present without compromising the ability of future generations to meet their own needs.” Regarding sustainable development, the UN [14] calls for “an integrated approach that takes into consideration environmental concerns along with economic development.”

Urban AI. Cugurullo et al. [9] refer to a “polymorphous agglomeration of AIs” encompassing “robots, autonomous vehicles, city brains and software agents” as urban AI.

Urban Sustainability. Chiu and Kikuzawa [15] describe urban sustainability as “the idea that a city is developed in a way that harmonizes economic growth, social inclusion, and environmental protection.”

1.2. Objectives

The primary objectives of this paper are to (a) provide a comprehensive review of the research and practice literature pertaining to urban sustainability, particularly in an era of AI; (b) identify key areas of focus and challenges emerging from the research and practice literature pertaining to urban sustainability in an era of AI; (c) identify gaps or problems in the research literature pertaining to urban sustainability in an era of AI; and (d) consideration of how AI specifically advances (or hinders) urban sustainability beyond high-level claims while being attentive to the reimagining and rethinking of urban and community sustainability initiatives emerging in this work as avenues for future directions going forward for research and practice in an era of AI.

1.3. Paper Overview

What follows is a description of the research and subsequent sections of the article beginning with the methodology (Section 2), a review of the research and practice literature (Section 3) encompassing urban and community sustainability (Section 3.1) and AI and urban and community sustainability (Section 3.2). Based on the literature review, Section 3.3 provide a conceptual mapping of the layers of urban and community sustainability and AI. Section 4 provides an overview of key focus areas and challenges for urban and community sustainability in an era of AI. A discussion of findings is provided in Section 5 in terms of gaps in Section 5.1 and opportunities for reimagining and rethinking urban and community sustainability aided by AI in Section 5.2, further probing of sustainability in Section 5.3, and recommendations in Section 5.4. Limitations of the paper and future directions for research and practice are identified in Section 6, followed by the conclusion in Section 7.

2. Methodology

This paper provides a comprehensive review of the research and practice literature, focusing on urban and community sustainability, particularly in an era of AI. Using Google Scholar, IEEE Xplore, Semantic Scholar, SpringerLink, and WorldCat, in support of an interdisciplinary review, literature was identified for the timeframe of 2020 to 2025. The primary search terms used were urban sustainability and AI. Using an analytic approach focusing on assessments, frameworks, gaps, and challenges within a limited timeframe (2020–2025) as filters, reduced the more than 6000 items identified, to approximately 200 items analyzed for this exploration. Research trends dominating the literature included the importance of frameworks for evaluation and assessment, collaborative initiatives and practices, multi-dimensional perspectives, and the need for ethical considerations going forward. Use of a community poll during a conference event (McKenna) [16] is also explored as a means of gathering urban sustainability insights from students, researchers, and practitioners, using the Whova online platform.

3. Literature Review

A review of the research and practice literature for urban and community sustainability is provided in Section 3.1 and in the context of AI in Section 3.2. From this literature review, key focus areas identified in Section 3.1 and Section 3.2 are then further probed, enabling a conceptual mapping in Section 3.3.

3.1. Urban and Community Sustainability

In early work on conceptualizing smart cities from an urban governance perspective, Nam and Pardo [11] identified the three dimensions of technology, people, and institutions, as well as the importance of sustainability. In a comprehensive study of urban sustainability indicators, Merino-Saum, Halla, Superti, Boesch, and Binder [17] identify three key tensions associated with the choosing of indicators, namely: parsimony vs. comprehensiveness, context-specificity vs. general comparability, and complexity vs. simplicity. As such, rather than indicators being purely technical, Merino-Saum et al. [17] place an emphasis on “the conceptual role that indicators play in (re)shaping the urban sustainability concept” as in, “making it tangible and operational in practice” pointing to the importance of social, need satisfaction, and resilience elements, to name a few. The importance of the use of frameworks or typologies in guiding the analysis of sustainable development is also advanced by Merino-Saum et al. [17] using the examples of SDGs (sustainable development goals) in combination with the STEEP (social, technological, economical, environmental, and political) and MONET (monitoring sustainable development) frameworks. From a health perspective, Crane, Simon, Haines, Ding, Hutchinson, Belesova, Davies, Osrin, Zimmermann, Capon, Wilkinson, and Turcu [18] advance a framework for transformative action economically, politically, and socially in rapidly meeting urban sustainability needs. Concerned with “unrealized potential” for the integrating of health and sustainability initiatives into urban life, the city of Copenhagen is used by Crane et al. [18] as an exemplar for municipal engagement with a variety of stakeholders “to integrate plans for climate mitigation and the promotion of health in everyday life.” Halla et al. [19] propose the use of metaphors, such as metabolism, rhythm, and smart, for understanding urban sustainability. The perspective offered by each metaphor as a cognitive tool highlights the importance of a multidimensional and transdisciplinary approach to the complexities of urban sustainability issues (Halla et al.) [19]. Hölscher and Frantzeskaki [20] explore urban transformation research using a structuring approach in relation to knowledge integration across the three perspectives, “in, of, and by cities”, for sustainability and resilience. Hölscher and Frantzeskaki [20] describe the practical aspects of the three transformation perspectives in terms of the ‘in’ as “collaborative place-making approaches like urban living labs” in support of local knowledge and the like; the ‘of’ in terms of “core urban strategies” using the examples of biodiversity and climate change; and the ‘by’ in terms of “policy knowledge exchange between cities.” Michalina, Mederly, Diefenbacher, and Held [21] provide a review of urban sustainability indicator frameworks (USIFs) for the measurement and assessment of sustainable urban development, identifying commonalities and differences, key dimensions, and primary indicators. Among key indicator frameworks highlighting assessment trends and challenges, in support of the work of policymakers, are those pertaining to water, mobility and transport, economy, waste, air quality, education, energy, land use, employment, health, housing, safety and security, and social infrastructure, to name a few (Michalina et al.) [21]. With an emphasis on solutions-oriented thinking, Wheeler and Rosan [22] seek to rethink unban sustainability using a community-oriented model focusing on economic and social issues, to name a few, using a social ecological lens and multi-disciplinary perspectives.

From a policy perspective, Bansard [23] explores paths to sustainable cities, highlighting the multi-dimensional and context-specific nature of urban sustainability, being attentive to urban-rural linkages, and circular resource use, to name a few. Gutierrez-Velez, Gilbert, Kinsey, and Behm [24] look beyond notions of the urban and rural in formulating “new infrastructure systems” supportive of rural-urban sustainability where nature is considered essential. Employing a futures lens, Chiu and Kikuzawa [15] address threats to urban sustainability focusing on governance implications. Sustainability-related issues are referred to by Chiu and Kikuzawa [15] as “wicked problems” as in, “highly complex, difficult to define, with multiple stakeholders and no immediate or obvious solutions.” Regenerative approaches to urban sustainability such as urban mining are highlighted by Chiu and Kikuzawa [15], involving “the recovery and reuse of a city’s materials from buildings, infrastructure, or products.” Providing an urgent and practical example, Zhang, Quoquab, and Mohammad [25] provide a knowledge mapping and analysis of research pertaining to plastics and sustainability, identifying directions for further research beyond the scope of their bibliometric analysis.

Yasuoka, Jensen, and Malmgren [26] advance the concept of “smart village” in identifying the importance of small communities in the context of future smart cities. Key features characterizing smart villages that are discussed by Yasuoka et al. [26] include participation, communication, resilience, infrastructure, and scalability. The importance of human relationships in smart environments is emphasized and the value of exploring sustainable initiatives in smart village living labs with the potential for the scaling of such initiative to the city level. Monoi, Kamio, Doi, Tamura, Kibi, and Yasuoka Jensen [27] advance the notion of an “-able city” model as an alternative to the efficiency and sustainability of the smart city concept enabled through digital technologies. By focusing more broadly on “citizen empowerment, participation and local resources to shape urban development” through the -able city lab (ACL), Monoi et al. [27] seek to involve people more actively in local initiatives in support of “more inclusive, participatory, and sustainable cities.”

According to the United Nations (UN) [28], urban growth is occurring for the most part “in small cities and intermediate towns” making more urgent the need for sustainable, resilient, and safe solutions. From this review of the research and practice literature of urban and community sustainability, key areas of focus that emerge include communication, culture, economy, environment (e.g., plastic pollution), governance, human relationships, infrastructure, regenerative approaches, resilience, small communities, socio-economic, and transport, to name a few. This would seem to be in keeping with the sustainable development goals (SDGs) for sustainable cities and communities as being concerned with some of the following elements (UN) [28] captured in Figure 1—cultural and social, disaster risk management (fire, floods, etc.), economy, energy consumption, environment (air pollution, heat, etc.), health and well-being, housing, infrastructure (transport, etc.), resilience, safety, and services.

Batty [29] explores cities in terms of histories, technologies, stories, and prediction where increasingly, AI and big data analytics are being used “to optimize transport, governance, infrastructure and sustainability” as explored in Section 3.2.

3.2. AI and Urban and Community Sustainability

The notion of artificially intelligent cities is advanced by Yigitcanlar and Cugurullo [1] along with the need for integrating AI into cities in sustainable ways. According to Yigitcanlar and Cugurullo [1], “the city of AI is not a sustainable city” and seven key issues requiring attention are identified: agility, ethics, monopoly, regulation, shareholders, social good, and trust. In response, Yigitcanlar and Cugurullo [1] share eight insights for improving the sustainability of AI in urban environments, from the combining of computational power and participatory planning to being prepared for AI disruptions to cities and societies, to name a few. Camaréna [30] formulates a practical, bottom-up “co-design process and agile approach” in support of sustainability where an Australian farmers market association could “experiment with an AI tool to link sustainable soil practices, nutrient rich produce, and human health” in a co-learning environment for iteratively considering ethical and sustainability questions involving transdisciplinary engagement. To further clarify, according to Hendawy and da Silva [31], top-down refers to a techno-centric approach, while bottom-up refers to a socio-centric approach, and in-between is a socio-technical approach. Nam and Pardo [11] acknowledge the success potential of both top-down and bottom-up approaches for smart city development, although they emphasize the importance of a “synergy” of active involvement from all sectors. From the perspective of urban planning, Jha, Ghimire, Thapa, Jha, and Raj [32] provide a review of the literature, focusing on the potential of use cases for AI and the Internet of Things (IoT) in the constructing of smart cities. While Jha et al. [32] claim that “everything has been changed by AI and IoT in smart cities” this would seem to be based more on such technologies having “promising effects on urban life” and research potentials going forward. In keeping with the UN SDGs for cities and communities, Koseki et al. [33] describe a Mila (Quebec Artificial Intelligence Institute)-UN-Habitat collaboration that uses people-centered and climate-sensitive approaches to considerations of risks, applications, and governance of AI and cities. Koseki et al. [33] advance a risk framework for the design, implementation, and maintenance of AI in cities; an urban AI strategy that is contextual, featuring cross-sectoral collaborations, capacity-building, and innovative regulatory tools for AI; and applications focusing on energy, mobility, public safety, water and waste management, healthcare, urban planning, and city governance. Schintler and McNeely [34] consider “prospects and provocations” for AI in relation to urban resilience, with implications for sustainability in terms of urban governance and planning. Significant challenges for AI and resilience highlighted by Schintler and McNeely [34] pertain to adaptive capacities and practices; the social effects of technology; ethics; and legal issues, to name a few. Zhang et al. [2] identify AI-related resources as data, knowledge, algorithms, and information systems with two types of orchestration for sustainable implementation including a policy-driven approach and an innovation-driven approach.

Koutalieris, Symeonidis, Kapsomenaki, Feio, Esposito, and Benis [35] describe the example of co-creative and agile methodologies for combining software development, environmental concerns, and health and well-being in support of urban environmental sustainability. Koutalieris et al. [35] describe the Unified Stakeholders Needs Co-creation Process (AENEA) as a “novel approach” to engagement involving “fluid, iterative collaboration and continuous feedback” that “clarifies, refines, and synchronizes stakeholder needs, fostering an open, and adaptable, collaboration framework for shared value generation.” It is worth noting that an AENEA implementation is being funded by the European Union (EU) for urban aquatic ecosystems (EU) [36] using “a novel AI-based surveillance system to monitor and protect” such spaces said to be important for “human mental and physical health.” Exploring the relationship between AI and urbanism from an urban studies and philosophy perspective, Palmini and Cugurullo [37] focus on how the application of AI technologies (e.g., robots, AVs, software agents) affects “urbanistic thinking and vice versa” in the interests of sustainable urban innovations. Palmini and Cugurullo [37] argue for a balanced approach to the use of technologies taking into consideration economic, environmental, and social factors in support of sustainable urban innovations, as in, “an ecology of intelligences.” From an architectural and urban design perspective, Hazbei and Cucuzzella [38] explore the complexities of context, noting that the concept consists of “tangible and non-human-related aspects (physical and environmental) that are amenable to quantification and measurement, on the one hand, as in, parametric design. On the other hand, Hazbei and Cucuzzella [38] identify the “intangible and human-related elements (cultural, social, and historical)” of the ‘context’ concept that are of a qualitative nature and less amenable to measurement. As such, Hazbei and Cucuzzella [38] point to future potentials for how the ‘context’ concept, in this fuller understanding, may serve to mediate sustainability outcomes and AI decision-making, through broader involvements and initiatives.

While Xu, Zhang, Gao, Feng, and Li [39] propose an urban generative intelligence (UGI) platform for agents in urban environments in the form of CityGPT, issues pertaining to responsibility, risk, privacy, and the like, do not seem to be addressed.

Shulajkovska, Smerkol, Noveski, and Gams [40] describe use of a “unified open-source simulation platform” prototype, tested in four European cities (Amsterdam, Bilbao, Helsinki, and Messina), in support of “more informed, data-supported decisions, improving urban mobility and sustainability” for policymakers involved in urban planning. As such, Shulajkovska et al. [40] advance an open-source AI framework for smart cities while identifying challenges such as that of system complexity which needs to be addressed to accommodate adoption and implementation, taking into account economic, environmental, and social factors. Tripathi, Bachmann, Brunner, Rizk, and Jodlbauer [41] undertake an assessment of the AI and sustainability literature, including the urban dimension, in search of trends, gaps, and challenges. Key research gaps identified by Tripathi et al. [41] pertain to an understanding of how to practically implement new technologies locally and “the use of AI within institutions, local businesses, healthcare, and education” encompassing issues of resource management, governance, and adverse impacts; whether and how AI integration on a global scale could occur; and the potential for the integrating of local and global concerns. Other areas highlighted by Tripathi et al. [41] include the importance of collaboration, ethical considerations, and the use of hybrid approaches, as in, “combining AI, data-driven techniques, and expert knowledge for multi-level, multi-dimensional decision-making.”

According to Batty [42], the open-ended nature of generative AI makes it particularly amenable to “designing solutions which improve the human condition” namely “the quality of life, the sustainability, and the prosperity associated with cities.” McKenna [43] calls for a rethinking of theory for urban life in an era of AI while offering a conceptual framework to guide the evolving, contesting, and understanding of urban theory. Theories included in the framework include ambient theory, assemblage theory and thinking, design theory, and theory of mind, to name a few, in support of adaptability, awareness, integration, literacies, and the like. Lartey and Law [44] critically address AI adoption in relation to urban planning governance, finding that while advancements would seem to be occurring in the development of AI capabilities for “predictive analytics and autonomous decision-making algorithms” an AI integration gap persists. That is, such AI advancements are said to be in conflict with the practical, “actionable application of AI in real-world governance” (Lartey and Law) [44] in that alignment issues arise with “regulatory, ethical, and participatory requirements” and ultimately with sustainability. It is perhaps worth noting that in a recent statement on AI and prosperity emerging from the Group of Seven (G7) countries forming an intergovernmental and economic forum (Prime Minister of Canada) [45], the expanding of “mutually beneficial partnerships with emerging markets and developing country partners to increase access to AI for everyone” is highlighted, involving initiatives such as the AI Hub for Sustainable Development by the United Nations Development Program (UNDP) [46], among others. Concerned with the sustainability of smart cities where people are centrally involved, Cezario, Moreira, Peres, Bilotta, Soares, and Guedes [47] explore how generative AI (Gen-AI) influences collaborative mapping implementations. Cezario et al. [47] “explore the influence of GenAI in the implementation of collaborative mapping with citizens,” said to be a “revolutionary method” where citizen participation supported by AI was found to “generate beneficial effects for reinforcing best sustainable practices” in keeping with the SDGs of the UN.

From the perspective of a human-centered AI (HCAI) approach, Herrmann [48] explores the potential of AI as a sparring partner (SP) in developing human capabilities in relation to learning, critical thinking, extending points of view, and creativity. Said to be valuable for people and teams, and extensible to computer-supported cooperative work (CSCW) (Herrmann) [48], it is possible that AI SP may have a valuable role for urban planning and sustainability. Worth noting is the call by Mittal [49] for a rethinking of approaches to distributed computing in an era of AI where “architectural innovation” is key and the need to “bridge the gap between distributed systems theory and AI practice” in “creating infrastructure that’s not just powerful but elegant.” In achieving the sustainability goals identified by the United Nations, aided by “responsible AI development and deployment,” Mutambara [50] identifies the need for visionary leadership across all sectors and organizational levels, extending to the global; addressing gaps in education; and garnering sufficient financial resources, among a series of other urgent requirements. Wang and Liu [51] explore the use of generative AI for cultural renewal initiatives focusing on the Chinese cities of Jinan, Wenzhou, and Hangzhou where artificial intelligence-generated content (AIGC) is currently in use. For example, the challenge of integrating cultural content with urban renewal is explored by Wang and Liu [51] “extensively drawing on and referencing the pioneering practices and valuable experiences of” such cities through use of GenAI, resulting in what is said to be an “innovative breakthrough in existing cultural models” with implications for “future urban development trajectories.” From an urban informatics perspective, Yue et al. [52] seek to address concerns with techno-solutionism by proposing a human–AI symbiosis framework for urban sustainable development.

From a business and practice perspective, Challapally, Pease, Raskar, and Chari [53] report that only 5% of 300 companies surveyed in the United States are experiencing success with GenAI initiatives. According to Challapally et al. [53], of the organizations that conducted GenAI evaluations (60%) only a small portion (20%) moved to a pilot phase, with fewer still reaching actual production (5%), and it is worth noting that failure was associated with “brittle workflows, lack of contextual learning, and misalignment with day-to-day operations.” On a practical level, Holmes [54] reports on the use of AI for road maintenance by one of many cities on Vancouver Island, Canada and the project is being conducted “in parallel with traditional road assessment practices.” According to Holmes [54], the 3-year trial project is showing promise economically in terms of cost for the value of data being generated and also in terms of integration “with other maintenance management systems” being used. It is perhaps also worth noting the text fragment forming part of the title, “reimagined road maintenance.”

Figure 2 provides a visual rendering for smart sustainable cities with AI integration in terms of key focus areas emerging from the literature review. Accessibility, creativity, the cultural, economy, education, and the environment emerge on the left with additional layering that includes innovation, social, and policy, while governance and infrastructure appear on the right in the realization of community and urban sustainability.

Underlying all components on the lower left is that of mapping and sensing and mobility and transport on the lower right. Also emerging from this review are key elements enabling the formulation of a conceptual mapping associated with the use of AI in urban environments as identified in Section 3.3.

3.3. Conceptual Mapping of Urban and Community Sustainability and AI

Based on this review of the research and practice literature a conceptual mapping is formulated in Figure 3, enabling an analytical surfacing of the four key thematic areas of approaches, contexts, frameworks, and practice (s) that are interwoven. The mapping further shows a layering of detail in relation to sustainable development goals, and then to challenges using the example of implementation, and to gaps using the example of education. Another layer of importance is that of impacts using the example of decision-making. With such layering, the mapping provides a deeper synthesis in and across domains. Where this paper is concerned mostly with SDG 11 focused on the urban, other works going forward could focus on all or any of the other SDGs (United Nations) [28].

Approaches addressed in the literature review include collaborative, integrative, and interdisciplinary while contexts of focus may be local or global. Of the many emerging frameworks, some include those for analysis, assessment, evaluation, and the identification of indicators for urban sustainability and AI. Where the focus is practice, the literature reveals the importance of being adaptive, collaborative, and cognizant of a range of perspectives.

4. Overview of the Focus and Challenges for Urban and Community Sustainability in an Era of AI

The primary areas of focus emerging from this review of the practice and research literature for urban and community sustainability in an era of AI are many and varied, as outlined in

Table 1. Focus and challenges for urban and community sustainability in an era of AI.

Areas of Focus	Frequency	Types of Challenges
Cultural	✓✓✓✓✓	Cultural content and urban renewal
Economic	✓✓✓✓✓✓✓✓✓✓✓✓	High costs of tech infrastructure
Environmental	✓✓✓✓✓✓✓✓✓✓✓	Water scarcity barriers
Ethical	✓✓✓✓✓✓✓✓	Alignment with urban planning
Governance	✓✓✓✓✓✓✓✓	Real-world actionable application of AI
Intelligence	✓✓✓✓✓	Agentic AI; human–AI
Privacy	✓✓✓✓✓	Data privacy and SLMs, STLMs
Regulatory	✓✓✓✓	Alignment; Outdated policies
Security	✓✓✓	AI model design and implementation
Social	✓✓✓✓✓✓✓✓✓✓✓✓✓	Social effects of technologies
Risks	✓✓✓	AI deployment
Technological	✓	Readiness
Trust	✓✓	AI systems

. Areas of focus include, but are not limited to, cultural, economic, environmental, ethical, governance, intelligence, privacy, regulatory, security, social, risks, echnological, and trustworthiness. From the 25 references cited in this paper with these areas of focus, the frequency of occurrence is indicated in Table 1 with a check mark. The social area of focus ranks highest with 13 of focus rank highest, each with checks, followed the economic with 12, the environmental with 11, governance and ethical each with 8 checks, cultural, intelligence, and privacy with 5, and regulatory with 4, and so on.

An example of the type of challenge for each area of focus associated with urban and community sustainability is also provided in Table 1 in the third column on the right.

For cultural, obstacles pertaining to the integrating of cultural content with urban renewal initiatives is considered and it is argued that generative AI would support “innovative explorations” (Wang and Liu) [51].

For economic, among the barriers to achieving urban sustainability according to Padi and Thangavelu [55] is that of costs associated with advanced technological infrastructure development.

For environmental, Padi and Thangavelu [55] highlight resource scarcity, such as water, in desert climates.

For ethical, Lartey and Law [44] note that AI adoption gives rise to challenges pertaining to alignment in the context of urban planning processes and policies.

For governance, the real-world actionable application of AI is highlighted by Lartey and Law [44].

For intelligence, challenges pertaining to agentic AI are highlighted by Tiwari [8] in terms of minimal, if any human oversight while the merging of human–AI intelligence in cities is presented as a grand sustainability challenge by Yigitcanlar and Cugurullo [1]. Indeed, Palmini and Cugurullo [37] question the appropriateness of the term intelligence for AI, arguing instead for the designation of agency rather than intelligence.

For privacy, concerns with urban data privacy (Weng et al.) [5] and metaverse platform data privacy (Owojori and Erasmus) [6] are highlighted, giving rise to the value of small language models (SLMs) in instances where privacy matters (Nanni, Chan, Lazauskas, and Geddes) [56] or possibly even super tiny language models (Sandhu) [57].

For regulatory, Lartey and Law [44] point to alignment issues with urban planning governance requirements while Koseki et al. [33] point to opportunities for developing innovative regulatory tools for AI, guided by a risk assessment framework.

For security, Weng et al. [5] identify security concerns with AI model design and implementation where “vulnerability to adversarial attacks and backdoor attacks” are acknowledged, affecting data security.

For social, Schintler and McNeely [34] explore whether and how AI may “hinder or compromise” urban resilience in relation to the social effects of technologies, emphasizing the need for “appropriate governance,” oversight, and human–AI collaborative arrangements.

For risks, Koseki et al. [33] present a framework for AI to address many types of risk, providing a guide for policymakers, encompassing as it does the five phases of the AI life cycle—framing, design, implementation, deployment, and maintenance.

For technological, Koseki et al. [33] identify barriers associated with technological readiness in relation to emerging technologies and the infrastructure to accommodate such technologies.

For trust, Yigitcanlar and Cugurullo [1] identify trust as one of several key elements that must be present for an AI city to be sustainable, and more specifically, “trust in AI systems” (Koseki et al.) [33].

In response to the research question posed in Section 1 of this paper, a proposition is provided here based on a reformulating of the research question, as follows:

P1. 

A rethinking and reimagining of sustainability is emerging in urban communities and regions in a variety of ways, particularly in an era of AI, in responding to a rich array of challenges and gaps.

This proposition is addressed in the discussion that follows in Section 5 in the context of findings from this review of the research and practice literature.

5. Discussion

What emerges from this review of the research and practice literature for urban and community sustainability in an era of AI is discussed in relation to gaps in Section 5.1 and to reimagining and rethinking initiatives in Section 5.2.

5.1. Gaps in the Literature for Urban and Community Sustainability in an Era of AI

Hölscher and Frantzeskaki [20] explore the interconnectedness of the three urban transformation perspectives of “in, of, and for” cities and how research gaps for urban sustainability and resilience may benefit from the complementary use of the three transformation perspectives, in support of awareness and learning. Leal Filho et al. [3] identify gaps related to addressing knowledge in support of sustainable practices; how AI might contribute effectively to multifaceted issues in cities; and research on using AI to promote sustainable cities. Cezario et al. [47] identify the gap in the research literature of understanding the impact of GenAI technologies on “the transparency and reliability of information” pertaining to urban and territorial sustainability. Herrmann [48] seeks to close the gap in the literature pertaining to the use of AI as a sparring partner in fostering learning and critical thinking. Huang, Bibri, and Keel [58] advance a large flow model (LFM) that is GenAI capable for integration with urban digital twins (UDT) systems to aid urban planning and design practices. According to Huang et al. [58], urban flows pertain to “mobility, goods, energy, waste, materials, and biodiversity” and are important for environmental sustainability when conducting modeling and analysis in support of “predictive analytics, adaptive learning, and complex data management functionalities.”

Mittal [49] identifies the need to bridge the gap between distributed systems theory and AI practice in achieving greater sustainability. Nanni et al. [56] address the public sector and compute-constrained environment gaps through use of small language models (SLMs), said to “perform remarkably” in environments “where consistency, speed, privacy, or cost matter more than state-of-the-art generality” and in support of sustainability concerns. While retrieval-augmented generation (RAG) aids SLMs, Nanni et al. [56] found that RAG alone is insufficient, such that with the use of synthetic examples, the lean model “was suddenly able to engage in thoughtful, grounded reasoning over real-world health content.” Sandhu [57] advances the notion of super tiny language models (STLMs) in support of “creating a safer, more sustainable AI ecosystem that benefits everyone.” In an opinion paper, Yue et al. [52] identify contextual challenges in relation to urban informatics giving rise to the need for “domain-specific AI innovations” and unified frameworks for modeling urban complexities.

Table 2. Gaps in the literature: urban and community sustainability in an era of AI.

Author(s)	Year	Gaps
Hölscher & Frantzeskaki [20]	2021	Urban transformation perspectives
Leal Filho et al. [3]	2024	Knowledge gaps for sustainable practices AI contributing effectively to multifaceted city issues Using AI to promote sustainable cities
Sandhu [57]	2024	STLMs for more sustainable AI ecosystems
Cezario et al. [47]	2025	Understanding the impact of GenAI
Herrmann [48]	2025	AI SP for learning and critical thinking
Huang et al. [58]	2025	Large Flow Model for UDT systems
Mittal [49]	2025	Distributed systems theory and AI practice
Nanni et al. [56]	2025	Public sector; Compute-constrained spaces
Yue et al. [52]	2025	Contextual challenges and urban informatics

provides an overview of gaps in the research and practice literature for sustainable cities and communities in an era of AI by author and year.

Using a visual approach, Figure 4 provides an overview of gaps pertaining to the use of AI for urban and community sustainability which may be interpreted as both challenges and opportunities. Such gaps pertain to distributed systems theory and the uses of AI, public sector challenges, knowledge in relation to sustainable practices, effective application of AI for urban challenges, the impact of GenAI, sustainable AI ecosystems, AI sparring partners for learning, large flow model for urban digital twins, and contextual challenges.

Further, such gaps and challenges are being addressed through the reimaging and rethinking of urban and community sustainability in an era of AI, as discussed in Section 5.2.

5.2. Reimaging and Rethinking Urban and Community Sustainability in an Era of AI

Yigitcanlar and Cugurullo [1] point to the need for a rethinking of “the economic dimension of cities” in an era of AI and, a reimagining of cities, as well, to “replan and redesign” such that “their function and shape are not dictated by and dependent on human economies.” Halla et al. [19] point to the potential for social and cultural change enabled through imagining new metaphors (e.g., smart, etc.) for more sustainable cities. In reimagining urban sustainability, Wheeler and Rosan [22] point to opportunities for action from “urban design to institutional restructuring” in accommodating change. Bansard [23] calls for a rethinking of how cities can be made liveable for everyone with minimal environmental impact. Schintler and McNeely [34] explore the potentials and challenges of AI for cities in relation to resilience thinking. Chiu and Kikuzawa [15] use a futures lens to rethink and inspire approaches to urban sustainability, advancing the notion of urban mining in recovery and reuse of materials as a regenerative approach. As if in confirmation of the work of Yigitcanlar and Cugurullo [1], Palmini and Cugurullo [37] point to an emerging understanding where “the future of the relationship between AI and the city passes through a rethinking of the economic” framework that guided “the smart city paradigm.” Lakusić et al. [4] undertake a reimagining of urban spaces in post-industrial European cities in developing sustainable futures focusing on demographic, educational, and environmental perspectives.

Kearns [59] provides a perspective on urban sustainability through the lens of learning cities calling for a rethinking of learning “as a driver of change” across the five domains of ecology, economy, community, well-being, and lifelong learning, extended to include culture, as in, the EcCoWell 3 approach. Monoi et al. [27] provide an alternative mindset and approach to urban development through the -able city notion involving the empowering of citizens, participation, and the use of local resources. And Mittal [49] calls for a rethinking of approaches to distributed computing with implications for sustainability. From an urban informatics perspective, Yue et al. [52] envision a future with technology and humanity evolving together, in the shaping of sustainable cities, involving a four-part symbiotic research agenda—adopting AI paradigms, modeling human perception, using generative AI responsibly, and enhancing human roles. And yet, however collaborative interactions are or have become with AI, as in human–AI collaboration, from a practice perspective, Mollick [60] claims that a move away from collaboration is occurring where humans become the “audience” as AI becomes more agentic.

Table 3. Reimagining and rethinking urban and community sustainability.

Author(s)	Year	Reimaging and Rethinking Urban Sustainability
Yigitcanlar & Cugurullo [1]	2020	Rethink economy and reimagine cities
Halla et al. [19]	2021	Metaphors for more sustainable cities
Wheeler & Rosan [22]	2021	Urban design and institutional restructuring
Bansard [23]	2022	Rethink livability with minimal environmental impact
Schintler & McNeely [34]	2022	Resilience thinking
Chiu and Kikuzawa [15]	2023	Futures lens: urban mining material recovery, reuse
Lakusić et al. [4]	2024	Reimagining post-industrial urban spaces
Kearns [59]	2025	EcCoWell 3 approach to sustainable futures
Monoi et al. [27]	2025	-able city as an alternative mindset and approach
Mittal [49]	2025	Rethinking approaches to distributed computing
Yue et al. [52]	2025	Envisioning mutual learning over automation

provides an overview of reimagining and rethinking initiatives in the research and practice literature for sustainable cities and communities in an era of AI by author and year, spanning the timeframe of 2020 to 2025.

Visually, Figure 5 identifies reimagining and rethinking initiatives for urban and community sustainability, with and without AI, which may also be interpreted as both challenges and opportunities. The visual includes the use of urban metaphors (e.g., metabolism, rhythms) for understanding the complexities of cities, urban design and institutional restructuring, mutual learning over automation, approaches to distributed computing and rethinking sustainability, -able city alternative mindset and approach, livability with minimal impact, resilience thinking, reimaging post-industrial spaces, and urban mining involving material recovery and reuse.

Such initiatives begin to provide insight into the presence and nature of emerging and evolving efforts for rethinking and reimagining sustainability in an era of AI that are manifesting in urban communities and regions in a variety of ways, in responding to gaps and challenges requiring urgent attention.

5.3. Probing Sustainability in an Era of AI

Drawing on the use of probes (Babbie) [61] and civic probes (Johnson and Vlachokyriakos) [62] encouraged in the research literature, this work employs a community online poll, during a live conference event, as a probe for learning more about sustainability and AI. The poll question (How might we rethink or reimagine sustainability in an era of AI?), options, and results are shown in

Table 4. Q: How might we rethink or reimagine sustainability in an era of AI?

#	Options	(% n = 5)
1	Engage with students on this	40%
2	City to city discussions	60%
3	Human–AI teaming	0%
4	I have no idea	0%
5	Other (please specify)	0%

.

While the number of responses to the poll question is small (n = 5), it is nevertheless worth noting that option 2—City to city discussions, received the highest response rate (60%) and this would seem to support a collaborative approach to sustainability challenges as advanced by Koseki et al. [33] and other researchers cited in this review paper. Equally of note is that no responses for option 3—Human–AI teaming, were received and this may speak to concerns with the state of AI in terms of trust (Yigitcanlar and Cugurullo) [1], ethics (Tripathi et al.) [41], and the like. That option 1—Engage with students on this received a 40% response rate is also of note and may point to the importance of education for AI and sustainability (Mutambara) [50], and new literacies for urban life in an era of AI (McKenna) [43].

5.4. Recommendations for Urban Sustainability Considerations in an Era of AI

From the exploration, analysis, and probing conducted in this paper a set of recommendations for urban sustainability considerations in an era of AI are provided, focusing on impacts, literacies, and policies.

Impacts. As shown in Figure 3, impacts are identified as one of several key layers in the mapping of urban sustainability and AI. As such, monitoring, assessing, and determining the outcome and effects of urban sustainability and AI initiatives to aid in the understanding of impacts will be a crucial area of research and practice going forward. Regarding impacts, Weng et al. [5] argue that AI security “will play an increasingly important role in shaping the future of digital, smart, and sustainable cities.”

Literacies. From an urban theory perspective, the development of existing and new literacies for AI is identified by McKenna [43]. From a practice perspective for AI, Mollick [60] calls for the “need to develop a new literacy” in terms of “when to summon” the use of AI, “when to work with AI as a co-intelligence” and when “to not use AI at all.” Given the rapid evolving of AI developments, new literacies will need to be more adaptive and dynamic.

Policies. As with new literacies in an era of AI, policies will need to be more adaptive and dynamic in support of rapidly evolving AI developments, implementations, and impacts. As such, AI education and new literacies for AI are crucial for policymakers, particularly if AI is being used to inform policy decisions (Koseki et al.) [33].

6. Limitations and Future Directions

Limitations of this work pertain to the ever-emergent nature of research and practice initiatives pertaining to urban and community sustainability such that, what may be offered in this paper is a glimpse or snapshot in time of what is occurring in a rapidly evolving era of AI. Another limitation of this work is the small sample (n = 5) for the conference poll response and this is mitigated perhaps by the nature of the exploration as a real-time, real-world probe for insights, in-the-moment. A range of sources are used in the literature review to mitigate the potential for bias and to strengthen the reliability of review findings.

Gaps identified in this review serve to provide future directions for research and practice—as in, knowledge gaps for sustainability practices in an era of AI, super tiny language models for more sustainable AI ecosystems, and the involvement of the public sector and compute-constrained spaces. Such gaps as future directions for research and practice would be guided by the recommendations in Section 5.4 focusing on impacts, literacies, and policies.

7. Conclusions

Among the contributions of this paper is the literature review, bringing together diverse perspectives, over the timeframe of 2020 to 2025, on urban sustainability in a rapidly evolving era of AI. Another key contribution of this work is a mapping of the layers of urban sustainability and AI, provided in Section 3.3, Figure 3, aiding as it does in sorting and understanding the complexities by thematic areas, SDGs, gaps and challenges, and impacts. Guided by the mapping, researchers and practitioners may choose one or more thematic areas (e.g., approaches, contexts, frameworks, practices), followed by one or more elements from the layers that follow, as they engage with urban sustainability and AI initiatives. The mapping is intended to be adaptive and may be extended and revised with use. Indeed, the mapping may enable articulations of new knowledge and new insights, through use.

New insights emerging from this paper are discussed in Section 5.3, based on a live conference poll probe of sustainability and AI, where confirmation of the importance of collaboration would seem to be suggested, an absence of support for human–AI teaming, and engagement with students that may be suggestive of support for education and new literacies for sustainability and AI. However, caution is encouraged in the interpretation of these findings, given the small sample size.

New knowledge emerging from this paper is presented in Section 5.4 in the form of recommendations for urban sustainability considerations in an era of AI, reflective of findings from the review of the research and practice literature. Focusing on the areas of impacts, literacies, and policy the recommendations are intended to be concrete and forward-looking for researchers and practitioners. This work is significant in that it confirms the presence of many initiatives to rethink and reimagine urban sustainability in response to the gaps and challenges presented by an era of rapidly evolving AI developments.

This paper will be of interest to researchers and practitioners concerned with urban design, governance, or planning; anyone concerned with urban digital twin design, development, implementation, or maintenance; and the informed and responsible integrating of AI, GenAI, or agentic AI into urban and community spaces.