Identifying Key Digital Enablers for Urban Carbon Reduction: A Strategy-Focused Study of AI, Big Data, and Blockchain Technologies

Abstract

The integration of artificial intelligence (AI), big data analytics, and blockchain technologies within the digital economy presents transformative opportunities for promoting low-carbon urban development. However, a systematic understanding of how these digital innovations influence urban carbon mitigation remains limited. This study addresses this gap by proposing two research questions (RQs): (1) What are the key success factors for artificial intelligence, big data, and blockchain in urban carbon emission reduction? (2) How do these technologies interact and support the transition to low-carbon cities? To answer these questions, the study employs a hybrid methodological framework combining the decision-making trial and evaluation laboratory (DEMATEL) and interpretive structural modeling (ISM) techniques. The data were collected through structured expert questionnaires, enabling the identification and hierarchical analysis of twelve critical success factors (CSFs). Grounded in sustainability transitions theory and institutional theory, the CSFs are categorized into three dimensions: (1) digital infrastructure and technological applications; (2) digital transformation of industry and economy; (3) sustainable urban governance. The results reveal that e-commerce and sustainable logistics, the adoption of the circular economy, and cross-sector collaboration are the most influential drivers of digital-enabled decarbonization, while foundational elements such as smart energy systems and digital infrastructure act as key enablers. The DEMATEL-ISM approach facilitates a system-level understanding of the causal relationships and strategic priorities among the CSFs, offering actionable insights for urban planners, policymakers, and stakeholders committed to sustainable digital transformation and carbon neutrality.

1. Introduction

Carbon emissions have become a pressing global issue [1,2]. In 2023, China emitted approximately 12.6 billion metric tons of carbon dioxide, accounting for 34% of the global total—the highest among all countries. According to the International Energy Agency, emissions continued to rise slightly in 2024, reaching an estimated 12.65 billion metric tons [3,4]. In response, the Chinese government implemented policies targeting peak emissions and carbon neutrality [5,6], aligning climate governance with high-quality economic development [7,8]. Consequently, low-carbon development has become a national priority, driving research into emissions pathways and influencing factors [9,10].

Recent scholarship increasingly recognizes AI, big data, and blockchain as key enablers in China’s transition toward low-carbon development, particularly within the broader context of the digital economy [11,12,13]. The integration of these emerging technologies—alongside Internet of Things applications—has proven instrumental in improving energy efficiency, promoting industrial upgrading, and enhancing urban sustainability governance frameworks [8,14,15]. China’s digital economy accounted for 38.6% of the national GDP, reflecting its central role in driving economic transformation and technological innovation [16,17,18]. This growing digital infrastructure lays a solid foundation for smart energy systems, intelligent transportation networks, and blockchain-based carbon trading mechanisms that are crucial to meeting national sustainability and decarbonization targets [19,20].

Key pillars of AI, big data, and blockchain-driven low-carbon urban transformation include digital infrastructure, intelligent energy management, sustainable logistics, and industrial digitalization [21]. The deployment of 5G networks, smart grids, and IoT-enabled sensors enhances urban carbon monitoring, allowing for real-time optimization of energy use and resource allocation [22,23]. AI-driven big data analytics and predictive modeling improve urban operations by enhancing energy efficiency in transportation, buildings, and industrial production [13,24]. The proliferation of e-commerce and digital logistics frameworks has precipitated the enhancement of sustainable supply chain management, thereby significantly diminishing carbon emissions associated with freight transportation [25,26]. Simultaneously, industrial digital transformation integrates smart manufacturing, predictive maintenance, and automated resource management, minimizing energy-intensive processes and reducing carbon footprints in key economic sectors [16,18]. As a result, AI, big data, and blockchain not only enable technological advancements in urban sustainability but also serve as strategic tools for achieving decarbonization goals [20,21].

Current research on the digital economy and carbon emissions primarily focuses on the relationship between digital technologies and carbon emissions, the mechanisms by which AI, big data, and blockchain reduce emissions, and carbon performance under various policy interventions [5,7,11,25]. However, studies often overlook the impacts of globalization, regional heterogeneity, diverse regulatory environments, and the longitudinal spillover effects of emerging technologies [1,2]. Additionally, integrative frameworks that combine environmental, economic, and institutional factors remain scarce.

The purpose of this study is to address these gaps by constructing an integrated framework for identifying the key success factors in urban carbon emission reduction through AI, big data, and blockchain within the digital economy context. The study explores the following research questions (RQs):

RQ1: 

What are the key success factors for artificial intelligence, big data, and blockchain in urban carbon emission reduction?

RQ2: 

How do these technologies interact and support the transition to low-carbon cities?

To address these gaps, this study applies sustainability transition theory and institutional theory to construct an integrated framework for identifying the key success factors in achieving urban carbon emission reduction through AI, big data, and blockchain within the digital economy context. This study adopts expert assessments and the DEMATEL-ISM method, integrating both qualitative and quantitative data collection methods. This approach enables a systematic analysis of the interactions among these key technologies, providing practical and policy-oriented insights for urban decarbonization [5,7,11,25].

This study analyzes China’s digital supply chain and financial services sectors to identify key determinants by which artificial intelligence, big data, and blockchain support urban carbon mitigation. Employing a tridimensional framework—digital infrastructure and technology deployment, digital economic transformation, and sustainable urban development—it integrates expert evaluations and the DEMATEL-ISM methodology, grounded in sustainability transitions theory and institutional theory. Sustainability transitions theory illuminates structural shifts toward low-carbon systems driven by digital innovations, while institutional theory examines regulatory, normative, and cognitive influences on technology adoption [27,28,29].

The findings highlight e-commerce, optimized sustainable logistics, circular economy practices, and cross-sector collaboration as critical factors influencing urban carbon reduction. Additionally, digital infrastructure, smart energy management, and intelligent transportation systems serve as foundational enablers. The DEMATEL-ISM method effectively clarifies complex interactions among these technologies, surpassing traditional regression methods through a nuanced systems-level analysis [27,28].

This research provides novel theoretical insights and practical strategies for urban planners, policymakers, and industry stakeholders, emphasizing AI, big data, and blockchain to enhance energy efficiency, carbon monitoring via digital finance and trading platforms, and governance through digital policy support, thus accelerating urban ecological sustainability. The goal is to equip stakeholders with technology-driven strategies that enhance cities’ global competitiveness in carbon mitigation and sustainable development initiatives.

The structure of the paper is as follows: Section 2 presents a synthesis of the relevant literature; Section 3 outlines the DEMATEL-ISM methodology; Section 4 reports the empirical findings; and Section 5 offers a comprehensive discussion and conclusion, including an acknowledgment of the study’s limitations and recommendations for future research directions.

2. Literature Review

2.1. The Role of AI, Big Data, and Blockchain in Urban Carbon Emission Reduction

The intensification of global climate change underscores the imperative of prioritizing energy conservation and emission reduction within the framework of international sustainability initiatives. The swift advancement of artificial intelligence, big data, and blockchain technologies within the digital economy provides innovative pathways for the progression of carbon emission reduction research [17,18]. Recent studies have significantly increased the exploration of the interplay between digital transformation and carbon emissions, utilizing a range of theoretical perspectives and methodological techniques [20,21]. However, significant discrepancies in research findings persist due to regional variations, sectoral differences, and methodological constraints. Empirical research indicates that advancements in the digital economy, notably in the realms of artificial intelligence, big data, and blockchain, are conducive to carbon reduction through enhanced energy management, increased industrial efficiency, and facilitated real-time emission monitoring [16,30,31]. Empirical evidence demonstrates that with the progression of the digital economy, its carbon mitigation impact intensifies, thereby substantiating the hypothesis of an inverted U-shaped relationship [22]. This suggests that initial digitalization may increase energy consumption, but over time, the efficiency gains from AI-driven optimization and blockchain-secured carbon tracking outweigh these effects, leading to a net reduction in emissions.

Nevertheless, some scholars contend that the expansion of digital infrastructure and data centers may result in increased energy consumption and associated emissions, particularly within energy-intensive domains such as cloud computing, artificial intelligence model training, and blockchain mining [16]. The ICT sector, despite driving digital innovation, remains reliant on energy-intensive operations, raising concerns about its net carbon impact. Moyer et al. [32] found that while AI, big data, and blockchain technologies can mitigate emissions, they may also introduce rebound effects by increasing demand for high-performance computing and large-scale data storage. This paradox underscores the need for more targeted policies that maximize digitalization’s benefits while mitigating unintended carbon impacts.

Regional disparities in carbon reduction effects further complicate the assessment of digital transformation’s environmental impact. Research has found that China’s eastern coastal regions experience more pronounced carbon reduction effects due to advanced AI adoption, blockchain-driven carbon markets, and industrial digitalization, whereas western regions with heavy reliance on fossil fuels see limited benefits [33]. The findings imply that the implementation of policy measures and regional adaptation strategies is imperative to harmonize the growth of the digital economy with sustainable carbon mitigation.

2.2. Methodological Approaches in AI, Big Data, and Blockchain-Driven Carbon Reduction Research

Investigations into the influence of the digital economy on carbon emissions have employed diverse methodologies, encompassing causal inference models, cointegration analysis, and threshold effect models. Cointegration analyses have revealed long-term equilibrium relationships among carbon emissions, digital transformation, and R&D investment, highlighting the structural interplay between technological advancement and sustainability [30]. Several scholars have utilized threshold and mediation effect models to investigate the moderating influences of industrial structure, urban scale, and resource endowment on the carbon reduction effects of the digital economy. Findings indicate that regions with high digital infrastructure investment experience stronger carbon mitigation effects, whereas areas with underdeveloped digital ecosystems face implementation challenges [15,22].

The mechanisms through which AI, big data, and blockchain contribute to emission reduction have been explored from multiple perspectives. Multiple studies have demonstrated that AI-driven predictive analytics, intelligent grid management, and blockchain-secured energy trading significantly enhance energy efficiency and contribute to emissions reduction [16,21]. The integration of artificial intelligence and the internet of things within industrial digitalization promotes energy optimization, reduces reliance on fossil fuels, and improves the efficiency of resource utilization. Moreover, scholars emphasize the pivotal role of industrial digitalization and e-commerce-enabled logistics optimization as key strategies for minimizing carbon footprints. AI-driven logistics route optimization, IoT-based intelligent supply chain management, and blockchain-assisted carbon footprint tracking synergistically boost the sustainability of urban transportation and industrial operations [18,34].

In spite of these advancements, current research frequently falls short in comprehensively capturing the intricate nature of the digital transformation’s influence on carbon emissions. Many studies focus on correlation rather than causation, leaving gaps in understanding the nonlinear and dynamic nature of digital technology adoption and its environmental effects [12,34]. Moreover, a scarcity of studies incorporates behavioral insights into the analysis of digital transformation, thereby circumscribing the breadth of research examining the impact of digital tools on consumer and corporate sustainability practices [13,35].

This study seeks to address these gaps by employing the DEMATEL-ISM methodology to elucidate the complex interdependencies among artificial intelligence, big data, blockchain, and urban carbon reduction strategies. Unlike traditional regression-based models, the DEMATEL-ISM approach facilitates the identification of causal linkages and hierarchical structures, thereby providing a comprehensive perspective on the long-term impacts of digital technologies on carbon mitigation.

2.3. Theoretical Underpinnings and Critical Success Variables

This study draws on sustainability transitions theory and institutional theory to explore the key success factors of AI, big data, and blockchain in urban carbon reduction.

Sustainability transitions theory provides a systemic and long-term perspective on how cities can shift from carbon-intensive to low-carbon socio-technical regimes. It underscores the dynamic interactions among niche innovations, established regimes, and broader landscape pressures [36,37]. Digital technologies, including artificial intelligence, big data, and blockchain, represent niche innovations that challenge established carbon-intensive systems through the facilitation of innovative approaches to energy management, transportation, and governance. These technologies expedite sustainability transitions through their disruptive capabilities—such as predictive energy optimization, decentralized carbon trading mechanisms, and real-time emissions monitoring—thereby transforming infrastructures and behavioral dynamics across diverse urban subsystems [36,37].

Meanwhile, institutional theory focuses on how organizational behaviors and policy systems evolve under formal regulations, normative expectations, and cognitive legitimacy [38,39]. Within the urban low-carbon framework, institutions exert influence on the adoption and dissemination of digital technologies via mechanisms including environmentally friendly regulatory frameworks, carbon disclosure standards, and digital governance models [39,40]. AI-powered carbon tracking platforms, blockchain-based compliance tools, and big data-supported policy evaluations reflect not only technological capability, but also the institutional embeddedness of sustainability initiatives [38,40]. The alignment of digital innovation with institutional forces is thus critical for ensuring long-term legitimacy, scalability, and policy effectiveness.

Drawing on sustainability transitions theory and institutional theory, this study identifies twelve pivotal success factors for urban carbon mitigation via ai, big data, and blockchain, grouped into three dimensions: digital infrastructure and technology applications (DITA), digital transformation of industry and the economy (DTIE), and sustainable development and green city construction (SDGCC).

The DITA dimension includes digital infrastructure, smart energy management systems, intelligent transportation systems, and data security in carbon tracking, representing niche-level innovations supporting energy and mobility decarbonization [27,41]. Institutional theory emphasizes regulatory trust and data governance, with blockchain enhancing transparency and accountability in urban climate governance. The DTIE dimension comprises industrial digital transformation, e-commerce and sustainable logistics optimization, circular economy adoption, and green digital finance and carbon trading platforms. These represent regime-level shifts in production, logistics, and finance catalyzed by AI and big data efficiency tools, driven by policy incentives, carbon pricing, and ESG norms [27,42]. The SDGCC dimension covers smart building and urban planning, digital governance and policy support, cross-sector collaboration and innovation, and citizen engagement and behavioral change. These reflect landscape-level shifts in urban planning, societal values, and governance models, integrating predictive AI and participatory platforms for enhanced sustainability and stakeholder collaboration [28,43,44].

By employing the DEMATEL-ISM methodology, this study systematically explores interactions among these factors, contributing theoretical insights and practical strategies for policymakers, businesses, and urban sustainability initiatives toward data-driven carbon reduction.

3. Methodology

3.1. Methodological Framework

Figure 1 illustrates the comprehensive research workflow employed in this study, followed by a detailed exposition of the DEMATEL-ISM analytical methodology.

A comprehensive literature review was undertaken across multiple academic databases, including Google Scholar, Science Net, SCOPUS, Web of Science, and CNKI, to identify twelve critical success factors influencing urban carbon emissions within the domains of artificial intelligence, big data, blockchain, and the digital economy. This review was supplemented by expert evaluations to guarantee the thoroughness of the factor selection. To ascertain the robustness and pertinence of the identified factors, qualitative interviews were conducted with three senior domain experts, each possessing over a decade of professional expertise in digital economy advancement, urban sustainability, and carbon mitigation policy. Utilizing the insights gained from these interviews, we conducted a thorough analysis of international regulations, standards, and guidelines relevant to urban carbon mitigation within the digital economy. By aligning our findings with sustainability transitions theory and institutional theory, we systematically assessed how these 12 key factors contribute to reducing urban carbon emissions. Sustainability transitions theory provided a framework to understand how digital technologies serve as enablers of systemic change in socio-technical urban regimes, while institutional theory offered insight into how policy frameworks, regulatory mechanisms, and social norms influence the adoption and implementation of these technologies. This evaluation process enabled us to identify, validate, and rank the primary factors that significantly impact carbon reduction efforts in smart cities, industrial digitalization, and urban sustainability (see

Table 1. Key success factors.

Code	Critical Success Factor	Definition	Sustainability Transitions Theory	Institutional Theory	Dimensions	Source
C1	Digital infrastructure development	Strong IT infrastructure (5G, cloud computing, IoT) is fundamental for enabling AI, big data, and blockchain applications in carbon reduction.	Support	Support	DITA	[1] [2] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [23] [24] [25] [27] [33] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50]
C2	Smart energy management systems	AI-driven smart grids and energy optimization enhance efficiency and integrate renewable energy into urban systems	Support	Support	DITA
C3	Intelligent transportation systems	AI and big data improve public transportation, reduce congestion, and optimize logistics for lower emissions	Support	Support	DITA
C4	Green digital finance and carbon trading platforms	Trading platforms blockchain enhances transparency and security in carbon trading, ensuring credibility in offset programs	Support	Support	DTIE
C5	industrial digital transformation	Smart manufacturing and predictive maintenance reduce waste and optimize resource use, cutting industrial emissions	Support	Support	DTIE
C6	Circular economy adoption	Digital tracking and AI-driven waste management facilitate resource recycling and sustainability in urban environments	Support	Support	DTIE
C7	Smart building and urban planning	AI-driven energy-efficient buildings and predictive urban development reduce emissions from construction and operations	Support	Support	SDGCC
C8	E-commerce and sustainable logistics	AI and blockchain optimize last-mile delivery, reduce emissions, and promote green supply chains	Support	Support	DTIE
C9	Digital governance and policy support	Data-driven policymaking ensures optimized regulatory frameworks for carbon reduction strategies	Support	Support	SDGCC
C10	Citizen engagement and behavioral change	AI-powered apps, gamification, and incentives encourage low-carbon lifestyles and sustainable consumer behavior	Support	Support	SDGCC
C11	Data security and privacy in carbon tracking	Blockchain secures emissions data, prevents fraud, and ensures transparency in climate reporting	Support	Support	DITA
C12	Cross-sector collaboration and innovation	Ecosystem partnerships between governments, businesses, and academia foster innovative low-carbon solutions through digital technologies	Support	Support	SDGCC

).

In the second phase of the study, we conducted a targeted expert selection and survey process to ensure that the evaluation of CSFs was grounded in high-quality, domain-specific knowledge. Our objective was to obtain informed judgments from professionals with deep experience in the application of AI, big data, and blockchain technologies for urban carbon reduction within China’s digital economy context. We began by identifying over 100 organizations operating across relevant sectors, including smart logistics, green energy, carbon trading, and industrial digital transformation. These organizations were sourced from academic publications, industry reports, and national digital economy project registries. After screening for relevance and expertise, 56 organizations were shortlisted. Each received a personalized email explaining the study objectives, data confidentiality policies, and the format of the expert engagement, which included a structured questionnaire and optional follow-up interview.

Out of the 56 shortlisted organizations, 27 responded positively to the invitation. However, due to scheduling and availability constraints, only 20 experts ultimately participated in the full survey and evaluation process. These 20 experts formed a highly qualified panel: all of them had over a decade of professional experience in digital technology and low-carbon urban development, and many held senior leadership or government advisory roles.

Table 2. Expert demographics.

Profile Details	Number of Respondents	Percentage (%)
Job Titles
Senior leadership	4	20
Digital economy experts	6	30
Academic scholars	4	20
Government policymakers	6	30
Years of Experience
10–15 years	4	20
16–20 years	12	60
≥21 years	4	20
Institution Type
Digital commerce enterprises	4	20
Low-carbon innovation enterprises	2	10
Green energy corporations	6	30
Logistics transportation companies	2	10
Environmental research institutions	6	30
Number of Employees
Fewer than 50	2	10
51–100	4	20
101–200	8	40
201–300	2	10
More than 301	4	20
Gender Distribution
Male	16	80
Female	4	20

Note: N = 20.

summarizes the demographic characteristics of the expert group, which included representatives from various sectors, specifically digital commerce enterprises (20%), low-carbon innovation enterprises (10%), green energy corporations (30%), logistics transportation companies (10%), and environmental research institutions (30%). In terms of job titles, 30% of the participants were digital economy experts, 30% were government policymakers, 20% were senior executives, and 20% were academic researchers. Notably, 60% of the panel had 16–20 years of experience, and 20% had over 21 years. This ensured that the panel could offer a mature, strategic perspective on both policy and technological implementation in the field of urban sustainability. Additionally, the gender distribution was 80% male and 20% female.

Initially, we piloted a traditional matrix-filling method, but several experts noted that aligning the codes with their respective definitions was time-consuming and prone to error. In response, we redesigned the instrument using a user-friendly Likert-style interface that dynamically displayed definitions and examples alongside each matrix cell.

All collected matrices were anonymized and aggregated to serve as the foundational input for the DEMATEL computation process.

Table 3. Original measurement table from one expert.

Factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12
C1	0	3	4	2	4	4	1	2	2	2	4	3
C2	2	0	4	1	3	1	3	4	0	3	1	4
C3	3	0	0	0	2	2	1	3	3	2	3	3
C4	0	2	4	0	2	4	0	1	3	0	3	1
C5	1	0	1	4	0	1	3	3	3	3	4	2
C6	0	3	1	3	1	0	1	3	4	1	1	3
C7	1	1	3	3	0	4	0	4	1	4	1	0
C8	3	3	3	4	0	4	4	0	0	0	0	0
C9	3	2	2	0	2	2	0	2	0	4	1	1
C10	0	3	0	3	1	0	4	2	3	0	2	2
C11	0	2	4	2	0	4	1	2	0	1	0	1
C12	1	4	2	0	3	4	3	4	4	2	4	0

Note: Each value represents the influence of the row factor on the column factor, as judged by an expert. The values range from 0 (no influence) to 4 (very strong influence). By definition, the diagonal values are set to 0.

presents the completed survey matrix from one expert as an illustrative example. The structured and systematic design of both the expert selection and the questionnaire development process ensures the robustness of the causal relationship analysis and supports the methodological rigor of the study.

In the third phase of this research, the DEMATEL-ISM model was employed to evaluate and deconstruct the critical success factors influencing urban carbon reduction within the framework of an AI-driven digital economy underpinned by big data and blockchain technologies. Urban carbon emission reduction represents a complex and systemic challenge, involving a wide array of technological, economic, and governance dimensions. While traditional quantitative methods like AHP and SEM provide valuable insights, they may inadequately capture intricate factor interdependencies (

Table 4. Comparison of the DEMATEL-ISM with other methods.

Method	Key Features	Strengths	Limitations
DEMATEL-ISM	Causal relationship analysis and hierarchical structure modeling	Effectively handles complex interdependencies and multi-layered issues	Relies on expert judgment, potentially introducing a degree of subjectivity into the analysis
Analytic hierarchy process (AHP)	Structured decision-making framework for ranking factors	Simple to use and suitable for multi-criteria decision-making	Struggles with interdependencies and complex causal relationships
Structural equation modeling (SEM)	Quantitative methodology for examining interrelations among multiple variables	Captures complex variable relationships and uses quantitative data	Requires large sample sizes and primarily identifies statistical associations rather than causality
Analytic network process (ANP)	Extension of AHP that incorporates interdependencies among factors	More effective than AHP in handling complex decision structures with feedback loops	Computationally intensive and requires expert judgment for weight assignment
Regression analysis	Statistical technique for modeling relationships between variables	Provides quantitative insights based on historical data	Assumes linearity and lacks the ability to model multi-layered or nonlinear interactions
Fuzzy logic	Decision-making method for managing uncertainty and imprecision	Adapted to accommodate ambiguous or imprecise data, beneficial for qualitative evaluations	Complex to implement and interpret, especially in large-scale systems

). In contrast, the DEMATEL-ISM model offers a more comprehensive analysis, effectively clarifying complex causal relationships and hierarchical structures [27,28,48].

Using the DEMATEL framework, we identified the interrelationships among key factors within AI, big data, and blockchain applications in carbon reduction, particularly in situations where data is incomplete or causal dependencies are unclear. The expert evaluation method played a crucial role in refining these interrelationships. DEMATEL uncovers the direct and indirect influences between factors, and through expert assessments, it constructs causal chains that help clarify how AI-driven digital infrastructure, blockchain-enabled carbon markets, and smart energy management contribute to urban carbon mitigation [27,49].

After determining the causal relationships of the key factors, we applied ISM to establish their hierarchical structure. ISM enables the stratification of complex systems, making the role and influence level of each factor more explicit [27]. This structured analysis allowed us to differentiate fundamental drivers of carbon reduction, such as digital infrastructure development and industrial digital transformation, from secondary effects, such as citizen engagement and digital governance policies. Identifying this hierarchy provides essential insights for developing targeted carbon reduction strategies and ensuring the effective integration of AI, big data, and blockchain in sustainable urban development.

Integrating the DEMATEL-ISM model into this study yielded several significant advantages. It revealed the complex interactions between digital technologies and urban carbon emission reduction through causal analysis, while hierarchical structure analysis clarified the roles and influence of each factor [50]. Compared to traditional methodologies, DEMATEL-ISM is particularly well-suited for assessing the impact of AI-driven analytics, blockchain-secured carbon tracking, and smart logistics in achieving urban decarbonization goals. The systematic methodology enabled a more nuanced evaluation of the digital economy’s role in urban carbon mitigation, further reinforcing its potential as a data-driven and policy-oriented mechanism for advancing sustainable urban development.

While recognizing the availability of certain objective data sources, such as carbon emission metrics, logistics performance indices, and digital infrastructure investments, this study employed expert-based evaluation due to the unique methodological and epistemological requirements of the DEMATEL-ISM framework. This research does not aim to quantify emissions directly, but rather to systematically explore the causal interdependencies and hierarchical relationships among strategic enablers of carbon reduction—relationships that are often nonlinear, qualitative, and context-dependent, particularly in emerging domains like AI-driven decarbonization.

Moreover, DEMATEL-ISM, as a method, is inherently reliant on expert judgment, especially when analyzing multi-dimensional and interaction-heavy systems where objective data are either sparse, inconsistent across regions, or insufficient to capture institutional and behavioral nuances. For example, factors such as digital governance, citizen engagement, and cross-sector collaboration are not readily represented by standardized datasets but are critical to urban decarbonization. The use of expert knowledge helps bridge these gaps, providing a structured yet flexible means of capturing system-wide influences.

To minimize subjectivity and enhance analytical robustness, the study incorporated rigorous control measures, including expert pre-screening (all with >10 years of experience), pilot testing, and a clearly defined Likert-scale questionnaire. As such, the expert judgment component complements rather than replaces empirical data, enabling a system-level causal analysis beyond the capabilities of conventional statistical techniques.

3.2. Research Approach and Techniques

Considering the multi-phased structure and computational intricacy of the DEMATEL-ISM methodology, the comprehensive process diagram is depicted in Figure 2, with an exhaustive breakdown of the procedural stages provided in the subsequent sections.

To systematically evaluate the interdependencies among the twelve identified success factors, this study employs a modified DEMATEL-ISM analytical process, which integrates influence mapping with hierarchical structuring. The computational procedure begins with constructing a pairwise influence matrix $Z=\left[z_{ij}\right]$, which quantifies the directional impact of factor $i$ on factor $j$, based on expert evaluations using a Likert scale. Notably, each diagonal element is set to zero ($z_{ii}=0$), excluding self-referential influence.

To ensure consistency in scaling and facilitate matrix convergence, the raw matrix $Z$ is transformed into a normalized influence matrix $N=\left[n_{ij}\right]$. The normalization is performed using the maximum of the row sums of $Z$, defined as

$$n_{ij}={\displaystyle \frac{z_{ij}}{\underset{i}{\mathrm{m}\mathrm{a}\mathrm{x}} {\sum }_{j=1}^n z_{ij}}}.$$

Following normalization, the cumulative interaction matrix $T=\left[t_{ij}\right]$ is derived. This matrix captures both direct and indirect influence pathways through an iterative convergence formulation based on matrix series expansion. The formulation is expressed as

$$T=N(I-N{)}^{-1},$$

where $I$ denotes the identity matrix. This transformation assumes that the spectral radius of $N$ is less than one, ensuring convergence. Each factor’s systemic characteristics are extracted from matrix $T$. The outgoing impact (or influence score) for factor $i$ is computed as

$$d_i=\sum _{j=1}^n t_{ij}.$$

The structural prominence of a factor or its centrality is given by

$$c_i=d_i+r_i,$$

and the directional orientation or causality degree is defined as

$$s_i=d_i-r_i.$$

The centrality scores $c_i$ are subsequently normalized to derive the relative weights $w_i$ of each factor in the overall system as follows:

$$w_i={\displaystyle \frac{c_i}{{\sum }_{i=1}^n c_i}}.$$

These metrics allow for the classification of each factor as either a driving (cause) or driven (effect) component in the urban carbon mitigation system.

To construct the reachability model, the enhanced influence matrix $H$ is obtained by incorporating self-influence into $T$:

$$H=T+I.$$

A predefined threshold $\theta$, typically set as the average of the elements in $T$, is applied to binarize $H$, resulting in the reachability matrix $R=\left[r_{ij}\right]$, where

$$r_{ij}=\left\{\begin{array}{ll}1,& h_{ij}\ge \theta \\ 0,& h_{ij}<\theta \end{array}\right..$$

From matrix $R$, the reachability set ${\mathcal{R}}_i$, antecedent set ${\mathcal{A}}_i$, and their intersection ${\mathcal{R}}_i\cap {\mathcal{A}}_i$ are computed for each factor. A factor is considered to belong to the top of a hierarchy layer if

$${\mathcal{R}}_i={\mathcal{R}}_i\cap {\mathcal{A}}_i.$$

Once identified, such factors are removed from the matrix, and the procedure is recursively repeated to establish the multilevel hierarchy of the system.

This combined influence-structuring approach provides a foundation for the causal mapping (centrality vs. causality) and layer-based system analysis, as demonstrated in the empirical findings. Compared to traditional statistical models, this method offers deeper insights into the functional architecture and influence dynamics of digital technologies in urban carbon reduction.

4. Results

After completing Steps 1–5,

Table 5. Comprehensive impact matrix.

Factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12
C1	0	0.09677	0.12903	0.06452	0.12903	0.12903	0.03226	0.06452	0.06452	0.06452	0.12903	0.09677
C2	0.06452	0	0.12903	0.03226	0.09677	0.03226	0.09677	0.12903	0	0.09677	0.03226	0.12903
C3	0.09677	0	0	0	0.06452	0.06452	0.03226	0.09677	0.09677	0.06452	0.09677	0.09677
C4	0	0.06452	0.12903	0	0.06452	0.12903	0	0.03226	0.09677	0	0.09677	0.03226
C5	0.03226	0	0.03226	0.12903	0	0.03226	0.09677	0.09677	0.09677	0.09677	0.12903	0.06452
C6	0	0.09677	0.03226	0.09677	0.03226	0	0.03226	0.09677	0.12903	0.03226	0.03226	0.09677
C7	0.03226	0.03226	0.09677	0.09677	0	0.12903	0	0.12903	0.03226	0.12903	0.03226	0
C8	0.09677	0.09677	0.09677	0.12903	0	0.12903	0.12903	0	0	0	0	0
C9	0.09677	0.06452	0.06452	0	0.06452	0.06452	0	0.06452	0	0.12903	0.03226	0.03226
C10	0	0.09677	0	0.09677	0.03226	0	0.12903	0.06452	0.09677	0	0.06452	0.06452
C11	0	0.06452	0.12903	0.06452	0	0.12903	0.03226	0.06452	0	0.03226	0	0.03226
C12	0.03226	0.12903	0.06452	0	0.09677	0.12903	0.09677	0.12903	0.12903	0.06452	0.12903	0

Note: Each value represents the total (direct + indirect) normalized influence of the row factor on the column factor. Values near 0 indicate minimal or no influence.

was generated, followed by

Table 6. Factor analysis outcomes: influence, centralization, and weighted orderings.

Factor	Influence Degree	Influenced Degree	Centrality	Causality	Weight	Ranking	Factor Attribute
C1	3.62739	1.771	5.39839	1.85639	0.08359	6	Causal Factor
C2	3.15938	2.74296	5.90234	0.41642	0.09139	5	Causal Factor
C3	2.64244	3.2621	5.90454	−0.61966	0.09142	4	Effect Factor
C4	2.308	2.63109	4.93909	−0.32309	0.07647	9	Effect Factor
C5	2.8122	2.05456	4.86676	0.75764	0.07535	10	Causal Factor
C6	2.49273	3.50656	5.99929	−1.01383	0.09289	2	Effect Factor
C7	2.49661	2.51352	5.01013	−0.01691	0.07757	8	Effect Factor
C8	2.53152	3.51665	6.04817	−0.98513	0.09365	1	Effect Factor
C9	2.3245	2.71036	5.03486	−0.38586	0.07796	7	Effect Factor
C10	2.33433	2.52576	4.86009	−0.19143	0.07525	11	Effect Factor
C11	1.9979	2.67825	4.67615	−0.68035	0.0724	12	Effect Factor
C12	3.56553	2.37971	5.94524	1.18582	0.09205	3	Causal Factor

Note. For “C” factor definitions, refer to Table 1.

, which presents factor analysis results of twelve critical success factors (C1 to C12) affecting urban carbon reduction via AI, big data, and blockchain. Table 6 includes each factor’s influence degree, influenced degree, centrality, causality, weight, rank, and classification as causal or effect factors. This analysis clarifies the hierarchical structure and interrelations among factors, aiding the identification of key components for optimizing urban carbon strategies and enhancing sustainability through AI, blockchain-based monitoring, and intelligent energy management.

Table 6 presents a multi-criteria evaluation of twelve key success factors (C1 to C12), using metrics including influence, dependence, centrality, causality, weight, ranking, and attribute classification. These metrics collectively address the first research question (RQ1), by identifying which factors play dominant roles in shaping AI, big data, and blockchain-enabled urban carbon reduction strategies. Specifically, influence captures a factor’s outgoing impact on others, while dependence reflects its vulnerability to external drivers. Their combination—centrality—offers insights into which factors act as systemic hubs. This aligns with sustainability transitions theory, wherein central actors often shape regime shifts through technological and institutional leverage.

The causality measure differentiates active drivers (cause factors) from passive outcomes (effect factors), directly addressing the second research question (RQ2) by illuminating the systemic logic behind urban carbon strategies. Weight reflects each factor’s normalized importance in achieving low-carbon outcomes, derived from its structural role and interconnections. The ranking of these factors—based on weight and centrality—prioritizes interventions, supporting evidence-based urban governance. The classification into cause-and-effect factors provides a stratified view of the digital transition process. This perspective is consistent with institutional theory, which emphasizes how actors and structures co-evolve, and supports the validation of our initial assumption: that digital infrastructure and governance mechanisms serve as enablers for downstream decarbonization outcomes.

These interdependencies are visually mapped in the causal relationship diagram (Figure 3), which plots centrality on the horizontal axis and causality on the vertical axis. This graphical representation allows for the identification of key levers (high-centrality, high-causality) and dependent outcomes (low-causality, high-dependence), aligning with strategic transition mapping frameworks. It offers not only a descriptive view but also a diagnostic tool for urban policymakers to identify high-impact digital levers in sustainability transitions.

The analysis indicates that e-commerce and sustainable logistics optimization possess the highest centrality value (6.04817) and weight (0.09365), highlighting their pivotal role within the carbon reduction strategies of the digital economy. Efficient e-commerce and logistics systems significantly lower transportation-related carbon emissions through AI-driven route optimization, real-time tracking, and green logistics. Blockchain technology enhances supply chain transparency, ensuring sustainable sourcing and reducing inefficiencies in freight and delivery systems. The results underscore the indispensable function of digital platforms in promoting low-carbon transportation and the decarbonization of urban freight systems. The adoption of the circular economy ranks second, with a centrality of 5.99929 and a weight of 0.09289. The circular economy model, enabled by AI-driven waste management, digital tracking systems, and blockchain-based material tracing, minimizes resource consumption, enhances recycling, and reduces urban waste. Although it exhibits a negative causality value (−1.01383), classifying it as an effect factor implies that its implementation is contingent upon other digital economy enablers, such as digital infrastructure development and cross-sector collaboration. Cross-sector collaboration and the innovation ecosystem rank third, with a centrality of 5.94524 and a weight of 0.09205. This factor highlights the importance of coordinated efforts between governments, businesses, and academia in leveraging AI, big data, and blockchain for urban planning, energy management, and industrial low-carbon transformation. Effective collaboration strengthens institutional frameworks, facilitates knowledge sharing, and accelerates the adoption of innovative technologies in sustainable urban development.

Figure 4 compares the centrality and weight of the top three factors, showing e-commerce and sustainable logistics optimization as the most influential, followed by the adoption of the circular economy, cross-sector collaboration, and innovation.

Digital infrastructure development ranks fourth, with a centrality of 5.39839 and a weight of 0.08359. Advanced 5G networks, cloud computing, IoT sensors, and edge computing provide the foundational support for AI-driven energy optimization, carbon monitoring, and smart city applications. A robust digital infrastructure ensures seamless data collection, real-time analysis, and automation of carbon reduction initiatives. Its positive causality degree (1.85639) indicates that it plays a critical enabling role in influencing multiple factors, reinforcing the need for strong digital transformation in achieving low-carbon cities. Smart energy management systems rank fifth, with a centrality of 5.90234 and a weight of 0.09139. AI-driven smart grids, demand-response systems, and energy storage optimization contribute significantly to improving urban energy efficiency. These technologies enable the seamless incorporation of renewable energy sources, thereby diminishing reliance on fossil fuels. The positive causality degree (0.41642) confirms that smart energy systems are an enabler of broader carbon reduction efforts, ensuring energy sustainability and resilience in urban environments. Intelligent transportation systems rank sixth (centrality = 5.90454; weight = 0.09142). Through smart traffic control, electric vehicle routing, and public transit analytics, they offer significant potential for emission reductions in the mobility sector. However, the negative causality score (−0.61966) positions this factor as a systemic outcome, rather than a driver. This reflects real-world dynamics where intelligent transport relies heavily on upstream digital infrastructure and urban policy environments. From an institutional theory lens, this illustrates how technological deployment in transport remains constrained by existing policy, funding, and planning structures, requiring coordinated institutional reform for scalable impact. Digital governance and policy support ranks seventh (centrality = 5.03486; weight = 0.07796), highlighting the role of institutions in facilitating carbon reduction through digital means. AI-assisted policymaking, blockchain-enhanced regulatory enforcement, and data transparency mechanisms promote trustworthy and adaptive governance. While its centrality is moderate, its qualitative influence is substantial, particularly in enabling coordination across sectors and ensuring the legitimacy of digital interventions. Smart buildings and urban planning (rank 8) contribute significantly to energy efficiency through AI-optimized systems and predictive maintenance. These solutions, when embedded into planning regulations and development incentives, improve the long-term energy performance of urban areas. Their placement in the mid-tier of influence suggests that while they offer measurable benefits, they depend on digital readiness and supportive governance. Green digital finance and carbon trading platforms (rank 9; centrality = 4.93909) reflect emerging financial instruments that support carbon-neutral economic shifts. Blockchain-backed carbon markets and AI-driven climate risk models facilitate transparent, traceable, and dynamic carbon pricing. These tools help operationalize the “market logic” of institutional decarbonization pathways, but their impact remains constrained without broader market adoption and regulatory maturity. Industrial digital transformation ranks tenth, focusing on AI- and IoT-enabled optimization within manufacturing, logistics, and utilities. Smart factories and predictive maintenance systems significantly improve energy efficiency and emissions performance in urban-industrial settings. While it holds moderate centrality (4.86676), this factor plays a vital role in achieving sectoral sustainability targets, particularly in developing economies with high industrial footprints. Citizen engagement and behavioral change ranks eleventh. Tools like AI-powered footprint apps, gamified incentives, and digital nudges promote sustainable consumer behavior. Although centrality and weight are relatively low, this factor is socially indispensable—public participation is essential for the legitimacy and scalability of digital climate strategies. From a normative institutional perspective, behavioral alignment represents the final link in translating digital innovation into sustained carbon impact. Data security and privacy in carbon tracking rank twelfth, with a centrality of 4.67615 and a weight of 0.0724. Blockchain technology provides tamper-proof carbon tracking, ensuring credibility in emissions reporting, carbon offset validation, and corporate sustainability claims. Secure and reliable carbon data management enhances trust in sustainability efforts and enables accurate measurement of carbon reduction progress.

The interrelationships among the twelve digital success factors were further examined using the DEMATEL-ISM methodology, with the results presented in

Table 7. Overall impact matrix.

Factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12
C1	1.14725	0.31594	0.38763	0.27897	0.29073	0.40723	0.24183	0.35654	0.29111	0.26979	0.34597	0.29442
C2	0.1939	1.19906	0.35193	0.2277	0.23765	0.28941	0.28558	0.37936	0.20193	0.27262	0.23049	0.28976
C3	0.20046	0.1756	1.19602	0.16352	0.18536	0.27796	0.18621	0.30208	0.25499	0.21409	0.25398	0.23217
C4	0.10008	0.20166	0.28938	1.13251	0.16946	0.3006	0.12728	0.21906	0.23768	0.13585	0.2303	0.16413
C5	0.14173	0.18208	0.24563	0.29248	1.12375	0.26634	0.2484	0.31011	0.2625	0.24824	0.28845	0.20248
C6	0.11006	0.2513	0.22111	0.23457	0.1517	1.20185	0.17648	0.2918	0.27274	0.17735	0.17961	0.22417
C7	0.13289	0.19304	0.27319	0.24759	0.11352	0.31793	1.14543	0.31527	0.19028	0.25308	0.17595	0.13843
C8	0.19345	0.24696	0.28844	0.2732	0.12516	0.3292	0.25605	1.20705	0.1628	0.14628	0.15596	0.14697
C9	0.18643	0.2108	0.2291	0.14665	0.17466	0.23914	0.14508	0.24813	1.14688	0.25493	0.17698	0.16571
C10	0.09791	0.23684	0.18442	0.23084	0.13753	0.19508	0.2551	0.2515	0.2269	1.14092	0.19693	0.18036
C11	0.08483	0.18633	0.26815	0.1794	0.09403	0.27977	0.14485	0.22466	0.13111	0.14006	1.11826	0.14647
C12	0.18202	0.34335	0.32709	0.22368	0.251	0.40205	0.30123	0.41109	0.33145	0.27254	0.32536	1.19466

Note: Each value indicates the overall influence of one factor on another, incorporating both direct and indirect effects. Higher values represent stronger total influence.

,

Table 8. Reachable matrix.

Factors	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12
C1	1	1	1	0	0	1	0	1	0	0	1	0
C2	0	1	1	0	0	0	0	1	0	0	0	0
C3	0	0	1	0	0	0	0	1	0	0	0	0
C4	0	0	0	1	0	1	0	0	0	0	0	0
C5	0	0	0	0	1	0	0	1	0	0	0	0
C6	0	0	0	0	0	1	0	0	0	0	0	0
C7	0	0	0	0	0	1	1	1	0	0	0	0
C8	0	0	0	0	0	1	0	1	0	0	0	0
C9	0	0	0	0	0	0	0	0	1	0	0	0
C10	0	0	0	0	0	0	0	0	0	1	0	0
C11	0	0	0	0	0	0	0	0	0	0	1	0
C12	0	1	1	0	0	1	1	1	1	0	1	1

Note: Binary values indicate whether one factor can reach another (1 = reachable, 0 = not reachable), based on a threshold applied to the overall impact matrix.

and

Table 9. Computing the reachability set, antecedent set, and set intersection.

Factors C	Reachability Set ${\mathit{R}}_{\mathit{i}}=\left\{{\mathit{f}}_{\mathit{i}}\right|{\mathit{F}}_{\mathit{i}\mathit{j}}=1\}$	Antecedent Set ${\mathit{S}}_{\mathit{i}}=\left\{{\mathit{f}}_{\mathit{i}}\right|{\mathit{F}}_{\mathit{j}\mathit{i}}=1\}$	Intersection ${\mathit{R}}_{\mathit{i}}\cap {\mathit{S}}_{\mathit{i}}$
C1	1, 2, 3, 6, 8, 11	1	1
C2	2, 3, 8	1, 2, 12	2
C3	3, 8	1, 2, 3, 12	3
C4	4, 6	4	4
C5	5, 8	5	5
C6	6	1, 4, 6, 7, 8, 12	6
C7	6, 7, 8	7, 12	7
C8	6, 8	1, 2, 3, 5, 7, 8, 12	8
C9	9	9, 12	9
C10	10	10	10
C11	11	1, 11, 12	11
C12	2, 3, 6, 7, 8, 9, 11, 12	12	12

Note: This table is used to determine the levels of factors in the structural hierarchy. Factors are grouped based on the comparison between their reachability and antecedent sets.

. Table 7, the overall impact matrix, quantifies both direct and indirect influence strengths among all factors, enabling a deeper understanding of systemic roles and validating RQ1 by identifying which factors exert the most significant influence across the network. Table 8, the reachable matrix, simplifies these relationships into binary form, clarifying which factors have the capacity to affect others either directly or through multi-step interactions. Table 9 builds on this by computing the reachability set, antecedent set, and intersection for each factor, providing the analytical foundation for hierarchical classification.

Together, these matrices support the verification of both RQ1 and RQ2: not only do they help pinpoint the most influential digital elements in urban carbon reduction strategies, but they also reveal the layered structure of their interdependencies. This structured output directly informs the development of the ISM framework in the next section, which visualizes how these factors align across different levels of influence, function, and strategic priority.

After applying the ISM method to perform a hierarchical analysis of the factors, the layered structure was obtained, as shown in Figure 5.

Based on the ISM hierarchical model (Figure 5), the analysis reveals a structured interplay among the 12 key success factors influencing urban carbon reduction through AI, big data, and blockchain-driven digital economy solutions. The hierarchical structure presents three key levels—fundamental factors, excessive factors, and surface factors—each playing a distinct role in enabling low-carbon urban transformation.

At the foundation level (L1), digital infrastructure development (C1) and cross-sector collaboration and innovation (C12) serve as the core structural drivers of urban carbon reduction strategies. These elements provide the essential technological backbone and institutional coordination required for the effective deployment of advanced digital solutions. Cross-sector collaboration fosters partnerships between government, businesses, and academia, ensuring that digital governance frameworks and technological innovations are well integrated with carbon reduction policy objectives. For example, urban testbeds in cities such as Singapore and Seoul demonstrate that investments in 5G infrastructure and collaborative innovation hubs are essential prerequisites for the large-scale deployment of AI-driven climate technologies.

Moving up to L2, smart energy management systems (C2) act as a critical bridge, linking infrastructure readiness to energy optimization strategies. Smart grid technologies, AI-enabled demand response systems, and blockchain-based carbon tracing tools play a vital role in enhancing urban energy efficiency, thereby supporting the integration of renewable energy within smart city frameworks. In practical applications, smart grid initiatives in cities like Copenhagen and Amsterdam have shown how AI-driven demand response mechanisms can reduce urban emissions while strengthening energy resilience.

The middle layer (L3) comprises industrial digital transformation (C5), intelligent transportation systems (C3), and smart building and urban planning (C7), which bridge technological advancements with urban operational efficiencies. Industrial digital transformation leverages AI and IoT to optimize energy-intensive industries, reducing emissions in manufacturing, logistics, and urban services. Meanwhile, intelligent transportation systems facilitate low-carbon mobility solutions, including EV adoption, AI-driven traffic management, and smart public transit systems, significantly lowering transport-related emissions. Smart buildings contribute to urban sustainability through AI-powered energy efficiency measures, promoting net-zero energy consumption in commercial and residential spaces.

At L4, e-commerce and sustainable logistics optimization (C8) and green digital finance and carbon trading platforms (C4) emerge as critical excessive factors that benefit from technological and policy advancements at the lower levels. E-commerce and logistics innovations powered by big data analytics reduce emissions from freight operations, while blockchain-secured carbon trading platforms ensure transparent, verifiable emissions offsets that facilitate a market-driven approach to carbon reduction.

The top level (L5) represents fundamental success factors, including the adoption of the circular economy (C6), citizen engagement and behavioral change (C10), digital governance and policy support (C9), and data security and privacy in carbon tracking (C11). These elements define the long-term sustainability outlook of urban decarbonization strategies. The adoption of the circular economy ensures that waste reduction, resource recycling, and sustainable consumption patterns are embedded within urban economic systems. Citizen engagement and behavioral change play a crucial role in shaping low-carbon consumer habits through AI-driven incentives, real-time carbon footprint tracking, and digital awareness programs. Digital governance mechanisms support policy enforcement, carbon tax frameworks, and regulatory compliance, while data security and privacy frameworks in carbon tracking ensure the integrity of emissions reporting and blockchain-enabled carbon audits.

The hierarchical structure ascertained through the DEMATEL-ISM model provides three key insights. Digital infrastructure development, smart energy management systems, and cross-sector collaboration and innovation serve as the core enablers that directly influence other carbon reduction strategies. Investment in smart grids, cloud computing, and AI-driven energy monitoring ensures long-term sustainability by optimizing energy distribution, transportation efficiency, and carbon accounting.

E-commerce and sustainable logistics optimization, the adoption of the circular economy, and intelligent transportation systems emerge as major outcome-driven factors that benefit from AI, blockchain, and big data integration. These success factors require robust policy frameworks, digital governance structures, and financial incentives to achieve scalable and impactful carbon reduction.

The interplay between smart building and urban planning, green digital finance and carbon trading platforms, and data security and privacy in carbon tracking highlights the need for a multi-layered approach that aligns digital infrastructure advancements with sustainable development goals. Strengthening policy frameworks, expanding digital governance mechanisms, and fostering innovation ecosystems will accelerate AI-driven policy innovations and carbon neutrality roadmaps for cities.

5. Discussion and Conclusions

This study explores the critical success factors for reducing urban carbon emissions in the context of the digital economy, using an integrated expert-based evaluation through the DEMATEL and ISM methodologies. By examining twelve key factors across three core dimensions—(1) digital infrastructure and technological applications; (2) digital transformation of industry and economy; (3) sustainable development and green urban construction—it establishes a coherent framework for understanding the complex interrelationships among these elements.

The analysis identifies several high-impact drivers of urban decarbonization enabled by the digital economy. Among them, e-commerce and sustainable logistics (C8), the adoption of the circular economy (C6), and cross-sector collaboration and innovation (C12) emerge as the most influential, underscoring their strategic importance in fostering systemic carbon reduction. These findings suggest that digital technologies must be supported by collaborative governance mechanisms and sustainable economic practices to achieve meaningful outcomes.

In addition, digital infrastructure development (C1) and smart energy management systems (C2) are confirmed as foundational enablers. They provide the essential technological backbone—such as 5G networks, cloud computing, IoT devices, and smart grids—required for real-time carbon monitoring, energy optimization, and data-driven urban services. Without these core infrastructures, the implementation of advanced carbon mitigation strategies would not be feasible.

Although blockchain technology holds significant promise for enhancing transparency, traceability, and efficiency in carbon markets, its broader application remains constrained by technical challenges, including high computational demands, limited scalability, and storage inefficiencies. Overcoming these barriers is critical to unlocking blockchain’s full potential in urban climate governance.

In sum, the findings reinforce the notion that a digitally enabled urban economy—driven by AI, Big Data, and blockchain—plays a vital role in advancing sustainable development and achieving carbon neutrality goals. Future research should continue to investigate the dynamic interplay among these technologies, with particular attention to implementation barriers and scalability. Building innovation ecosystems and strengthening policy frameworks will be essential for cities seeking to leverage digital transformation in their low-carbon transitions.

Importantly, the proposed DEMATEL-ISM framework, although designed specifically for urban carbon reduction within the digital economy, demonstrates strong adaptability across other policy domains and industrial sectors characterized by complex interdependencies. It can be applied to strategic planning in areas such as smart agriculture, green manufacturing, and regional decarbonization logistics, provided the factor set and expert inputs are tailored to local conditions. This adaptability highlights the model’s value in both academic inquiry and practical policy analysis.

Ultimately, the framework serves as a flexible decision-support tool for planners and policymakers aiming to align technological innovation with sustainability objectives. However, its effective application requires thoughtful contextualization and active stakeholder engagement to ensure relevance, legitimacy, and impact across diverse urban and sectoral settings.

5.1. Theoretical Repercussions

This study provides timely and substantial contributions by offering novel theoretical insights into the role of artificial intelligence, big data, and blockchain in advancing low-carbon urban transformation within the digital economy.

First, it extends the theoretical framework for understanding the optimization of urban carbon reduction through digital technologies by systematically examining the key success factors that underpin sustainable urban development. While the extant literature has examined the general effects of the digital economy on carbon emissions, a limited number of studies have conducted a comprehensive analysis of the specific mechanisms through which digital transformation factors, including smart energy management systems, intelligent transportation systems, and digital infrastructure development, contribute to urban decarbonization [51,52]. By constructing an analytical framework integrating sustainability transitions theory and institutional theory, this study provides a comprehensive evaluation of how AI-driven energy optimization, Blockchain-secured carbon tracking, and Big Data-enabled governance models contribute to carbon reduction strategies. Furthermore, through the identification and hierarchical analysis of 12 key factors, this research builds upon and extends the conclusions of Ma et al. [53], bridging theoretical gaps and enhancing the practical application of AI, big data, and blockchain in urban sustainability initiatives.

Secondly, this research contributes to the comprehension of the economic, environmental, and policy-related advantages linked to the deployment of artificial intelligence, big data, and blockchain for reducing urban carbon emissions. Empirical findings reveal the manner in which digital technologies facilitate energy efficiency, industrial productivity enhancement, advanced carbon management, and the encouragement of eco-friendly behavioral changes. Notably, e-commerce and the optimization of sustainable logistics, alongside the adoption of circular economy principles, and the development of digital infrastructure were pinpointed as the most pivotal catalysts for fostering a low-carbon urban transformation [53,54]. These findings offer robust theoretical foundations to support urban administrators, policymakers, and business leaders in formulating data-driven and technology-enabled strategies for effective urban decarbonization. Additionally, this study reinforces the importance of blockchain-enabled green digital finance and carbon (finance and carbon) trading platforms in ensuring transparent carbon markets, incentivizing sustainable investments, and strengthening carbon accountability mechanisms. In contrast to prior research, including that of Moyer and Hughes [32], which centered on macroeconomic trends, the present study employs a multi-faceted methodology that synthesizes expert assessments with the DEMATEL-ISM framework. This approach is employed to systematically analyze the complex interrelationships between technologies driven by the digital economy and urban carbon mitigation policies.

Ultimately, this research enriches the theoretical and methodological synthesis of digital transformation with sustainable urban development. By synthesizing sustainability transitions theory with institutional theory, it identifies the key success factors through which artificial intelligence, big data, and blockchain facilitate digitally driven carbon reduction. Utilizing a combined quantitative and qualitative analytical framework, the study delves into the causal linkages and hierarchical configurations among elements like smart buildings and urban planning, digital governance and policy advocacy, as well as intersectoral collaboration and innovation. This exploration yields theoretical perspectives and actionable strategies for policymakers and urban sustainability professionals. The proposed research framework serves as a benchmark model for subsequent studies, offering a systematic approach to understanding how emerging digital technologies can further accelerate low-carbon urban transformation and carbon neutrality goals. By integrating AI-powered predictive analytics, blockchain-driven carbon accountability, and big data-enabled smart policy interventions, this research lays the foundation for scalable, technology-driven solutions that support long-term environmental sustainability.

5.2. Strategic Implementation Framework

For the effective abatement of urban carbon emissions, it is imperative that city administrators, governmental bodies, and corporate stakeholders collaborate to execute comprehensive strategies. These strategies must harness the power of artificial intelligence, big data analytics, and blockchain technology within the context of the digital economic model. The following strategic recommendations delineate how key stakeholders can optimize the transformative potential of these technologies to advance low-carbon urban development, in alignment with the twelve key success factors identified in this study. Initially, municipal administrators should formulate a strategic framework focused on sustainable digital development. This framework should prioritize the enhancement of digital infrastructure, the adoption of intelligent energy management systems, and the integration of smart transportation systems, all as essential steps for diminishing urban carbon emissions [52]. Investment in 5G networks, cloud computing, and IoT sensors is essential for enabling real-time carbon tracking, AI-driven smart grid optimization, and efficient urban transportation systems [44]. AI-powered smart buildings and urban planning can further optimize energy use, minimize resource wastage, and enhance overall urban energy efficiency. Additionally, integrating blockchain-based green digital finance and carbon trading platforms will facilitate transparent, verifiable carbon offset mechanisms and secure emissions reporting. A successful low-carbon digital transformation must also consider social and economic welfare, ensuring that digitalization projects align with sustainability standards and generate long-term environmental benefits [53].

Second, government agencies play a pivotal role in regulating and guiding digital economy-driven carbon reduction initiatives. Establishing comprehensive policy frameworks that support AI-powered carbon monitoring, Smart energy management systems, and Blockchain-enabled carbon credit mechanisms can accelerate urban sustainability efforts. Governments are recommended to provide tax incentives, green financing initiatives, and financial subsidies to incentivize enterprises to allocate resources towards e-commerce and sustainable logistics optimization, the adoption of circular economy principles, and the implementation of energy-efficient industrial transformations [51]. Additionally, fostering international cooperation in AI-driven climate governance and blockchain-enabled carbon markets will enhance global carbon reduction efforts. Governments should also leverage digital governance and policy support to implement robust carbon accountability frameworks and launch public awareness campaigns that promote citizen engagement in low-carbon urban initiatives. Encouraging citizen engagement and behavioral change through AI-driven sustainability incentives and digital carbon tracking platforms can empower individuals to make climate-conscious decisions in urban mobility, consumption, and energy use.

Finally, businesses are integral to incorporating AI, big data, and blockchain into industrial digital transformation to reduce carbon emissions. Enterprises should embed green digital transformation into their core strategies, leveraging AI-driven energy analytics, IoT-enabled smart manufacturing, and blockchain-secured supply chain transparency to reduce emissions in industrial production. Optimizing logistics and transportation systems through AI-powered predictive analytics and blockchain-enabled tracking can significantly cut fuel consumption and minimize environmental impact [45]. Additionally, companies should invest in AI-powered smart energy management systems, distributed renewable energy solutions, and digital carbon footprint monitoring to advance sustainability in production. Employee training programs in AI-driven energy efficiency, big data-based carbon tracking, and blockchain security can enhance workforce competencies and strengthen corporate sustainability leadership [33].

By implementing these integrated strategies, city administrators, government agencies, and businesses can collaboratively address urban carbon challenges, maximize the potential of AI, big data, and blockchain, and accelerate the global transition toward low-carbon, sustainable cities. Through strategic investments in smart energy management, the adoption of the circular economy, and digital governance and policy support, cities can leverage cutting-edge digital solutions to drive carbon neutrality efforts and long-term urban sustainability.

5.3. Limitations and Future Research

Despite the comprehensive nature of the DEMATEL-ISM analysis, several limitations must be acknowledged. First, the reliance on expert judgment may introduce subjective biases, which may affect the reliability and generalizability of the identified causal relationships. Second, the study’s exclusive focus on China’s digital economy may limit the applicability of its findings to regions with different regulatory environments or levels of technological maturity. Additionally, the analysis does not fully capture long-term behavioral dynamics or the heterogeneous impacts of digital technologies across diverse urban contexts. A further limitation lies in the still-nascent application of blockchain technology, despite its considerable potential in carbon tracking and digital finance. The widespread adoption of blockchain remains constrained by challenges related to computing power, data storage, and integration with existing infrastructure, limiting its practical deployment in urban carbon management. As such, the study’s discussion of blockchain may reflect theoretical expectations more than established real-world practices. To address these limitations, future research should incorporate more diverse datasets, integrate expert evaluations with empirical data, and apply advanced analytical methods to enhance validity. Moreover, dynamic system modeling and cross-regional comparative studies could further illuminate how AI, big data, and blockchain technologies contribute to low-carbon urban transformation. In particular, future work should examine scalable, real-world applications of blockchain in this domain.