How Does Artificial Intelligence Drive the Optimization of Public Services Structural? ——Complex Intermediary Mechanisms Based on the Digital Ecosystem

How Does Artificial Intelligence Drive the Optimization of Public Services Structural? ——Complex Intermediary Mechanisms Based on the Digital Ecosystem | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article How Does Artificial Intelligence Drive the Optimization of Public Services Structural? ——Complex Intermediary Mechanisms Based on the Digital Ecosystem Chunlin Xiong, Ren Fan, Yi Liu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8350814/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Against the backdrop of the "Digital China" strategy, the digital ecosystem has emerged as a strategic fulcrum for transforming public service delivery from a supply-oriented to a demand-oriented paradigm, with artificial intelligence (AI) serving as the core engine of structural optimization of public services. However, existing research has rarely examined the complex causal mechanisms by which AI, through configurations of digital ecosystem elements, shapes the structural optimization of public services. Based on panel data from 30 Chinese provinces covering 2019–2023, this study integrates the configurational perspective with digital ecosystem theory to develop a systematic“technology input–system integration–value output”framework. Employing an advanced mediation model that combines dynamic qualitative comparative analysis (QCA) and regression methods, the research elucidates the pathways and process mechanisms by which AI drives the structural optimization of public services. The findings indicate that AI significantly promotes the structural optimization of public services, achieving multi-level upgrades through three distinct configurations of the digital ecosystem: (1) ecologically mature type (an advanced, all-encompassing development model), (2) administratively driven type (a government-led approach), and (3) socially empowered type (an inclusive development pathway).This study clarifies the mechanism by which AI and the digital ecosystem collaboratively reshape the architecture of public services, highlighting that achieving public service modernization requires not only strengthening he development and application of AI technologies but also fostering a compatible digital ecosystem. Such synergy creates a virtuous cycle of technological empowerment and systemic optimization, ultimately maximizing public value. The results provide empirical evidence and forward-looking insights for inclusive, digitally integrated governance in the era of intelligent technologies. JEL: H83; O33 Business and commerce/Information systems and information technology Social science/Science technology and society Artificial Intelligence Optimization of Public Service Structure Digital Ecosystem Complex Mediation Dynamic QCA Figures Figure 1 Figure 2 Figure 3 Introduction In recent years, the rapid global advancement of artificial intelligence (AI) technology has catalyzed profound transformations in public service delivery. Recognizing AI's strategic importance, numerous countries have accelerated the formulation and implementation of dedicated policies. In 2024, the European Union adopted the Artificial Intelligence Act, establishing foundational regulatory safeguards for AI applications in high-risk domains such as public administration. According to the Organization for Economic Cooperation and Development (OECD), over 70% of its member states have integrated AI into their digital government strategies, yielding significant improvements in critical public service sectors including healthcare, education, and transportation. In China, the "AI+Public Services" initiative was formally proposed in the 2025 Guidelines on Deepening the Implementation of the "AI+" Initiative. As the "Digital China" strategy continues to advance, AI—positioned as a pivotal driver of the new wave of technological revolution and industrial transformation—plays an increasingly central role in addressing systemic challenges in public services, including insufficient coordination 1 , technical adaptation barriers 2 , information asymmetry 3 , and unequal resource distribution 4,5 . However, empirical evidence indicates that isolated, single-purpose applications of AI technology can achieve only localized efficiency gains and are largely insufficient to drive systemic and holistic quality improvements in public service systems 6 . These limitations are manifested across three primary dimensions: First, the pervasive existence of "data silos" impedes cross-departmental and cross-level data sharing, thereby undermining the integrated effectiveness of AI deployment. Second, a persistent misalignment between technological research and development (R&D) and actual public service demands results in solutions that fail to align precisely with real-world contexts, weakening their capacity to generate tangible value. Third, a critical shortage of interdisciplinary talent—professionals equipped with expertise in both AI and public administration—has emerged as a key bottleneck constraining the deep integration and sustainable application of digital technologies in governance. Thus, the essential catalyst for transforming public services is the construction of a synergistic and high-performance digital ecosystem, premised on the deep embedding of technology within practical operational contexts. This foundation supports the progressive maturation of applications across critical public service domains, including government approvals, smart healthcare, and accessible education. Consequently, the pivotal issues demanding immediate scholarly and policy attention revolve around elucidating the pathways through which artificial intelligence drives the optimization of public service frameworks, and defining the architecture of a digital ecosystem capable of fully activating AI's potential, overcoming the "digital divide," and thereby realizing structural optimization in public services. In light of this, the present study combines configuration theory with digital ecosystem theory, employing a hybrid approach of dynamic QCA and regression analysis. A sample of 30 provinces in China (excluding Tibet) from 2019 to 2023 was utilized to overcome the limitations of traditional research frameworks that exclusively prioritize technology application. Concurrently, a systematic analytical framework encompassing five dimensions—innovation, green development, openness, sharing, and coordination—is being constructed to explore the complex causal mechanisms through which AI influences public service structural optimization within the combination of digital ecosystem elements. This provides empirical evidence or theoretical frameworks and practical insights for the digital and intelligent development of public services. This study makes significant marginal contributions theoretically and methodologically. Theoretically, it introduces an innovative perspective that synthesizes AI's direct influences with a nuanced analysis of the digital ecosystem's mediating functions, thereby challenging the prevalent technological determinism in public administration scholarship. Methodologically, it pioneers a novel approach by synthesizing dynamic QCA with complex mediation analysis. This synthesis adeptly elucidates the multiple concurrent causal configurations whereby AI impacts public services through the digital ecosystem, furnishing a novel analytical framework for interrogating complex causal mechanisms in social sciences. Literature Review 1.1 Research on Artificial Intelligence and Public Services In recent years, the application of artificial intelligence (AI) to transform and enhance public services has become a prominent focus in academic research. This body of scholarship has primarily centered on three interrelated domains: AI applications in public service delivery, the transformation of government governance models through AI, and AI-enabled mechanisms for public feedback. First, artificial intelligence and public service applications. Current scholarly research primarily treats AI as a technological enabler applied across specific service domains, organized around three key dimensions. First, intelligent interaction and automation 7 : generative AI and chatbots are widely deployed in government hotlines, tax consultation, legal aid, and related services, enabling 24/7 digital responsiveness and significantly improving service accessibility and response speed 8 . Second, data-driven decision-making: big data analytics and machine learning algorithms are leveraged to anticipate public demand and optimize resource allocation 9,10 . For example, in smart healthcare, AI supports clinical diagnosis and epidemic forecasting; in smart transportation, it facilitates real-time traffic management and accident prediction 11 . Third, technology-enabled empowerment and security: emerging technologies such as blockchain enhance the transparency, traceability, and integrity of government data 4 . In critical areas including welfare distribution, government procurement, and identity verification, these tools help strengthen public trust 12 . Collectively, these AI applications not only streamline service delivery processes but also reduce administrative burdens 13 , thereby enhancing both operational efficiency and service quality. Second, artificial intelligence and the transformation of government governance models. The application of artificial intelligence has profoundly reshaped the organizational logic and operational patterns of governments, accelerating the modernization of governance systems. First, it drives a shift in organizational structures from traditional bureaucratic hierarchies toward flat, networked, and data-driven configurations. This transition necessitates dismantling entrenched data silos to enable efficient cross-departmental collaboration 14 . Second, it enhances scientific decision-making 15 . AI provides robust capabilities in simulation, scenario analysis, and impact assessment for policy formulation, facilitating a transition from experience-based judgment to data-driven, precision-oriented analysis 16 . Third, it fosters the emergence of a new ecosystem of multi-stakeholder collaborative governance, in which governments, enterprises, social organizations, and citizens form synergistic partnerships through AI-powered platforms 17 . However, this transformation also presents significant ethical challenges—including algorithmic fairness, accountability mechanisms, and data privacy—that demand the urgent development of comprehensive governance frameworks 18 . Third, AI and public feedback. The ultimate effectiveness of AI applications is reflected in citizens’ sense of fulfillment, satisfaction, and the generation of social value. On one hand, AI-driven personalized services can precisely align with citizens’ diverse needs 19 , thereby significantly improving public service experiences and user satisfaction 20 . On the other hand, careful attention must be paid to the risks posed by the "digital divide" 21 . In the absence of inclusive design, AI applications may marginalize vulnerable groups such as the elderly and persons with disabilities, potentially exacerbating existing social inequalities. Therefore, advancing inclusive design requires ensuring that technological progress achieves genuine universality, enabling equitable access and participation—conditions essential for translating innovation into enhanced social welfare and strengthened government credibility 22 , which constitutes the core of public value creation. 1.2 The Systematic Evolution of Digital Technology into a Digital Ecosystem In recent years, academic inquiry has predominantly focused on the application pathways of artificial intelligence in public services 23,24 . However, much of this scholarship has treated AI as a homogeneous technological instrument, emphasizing its general utility and outcomes—resulting in fragmented rather than systematic solutions. A deeper analysis reveals a significant evolution: digital technologies are increasingly coalescing into an integrated digital ecosystem. This transformation redefines technology’s role from a mere tool for enhancing efficiency to a structural force capable of fundamentally reshaping patterns of production and daily life, models of governance, and even broader civilizational frameworks. Consequently, this shift extends beyond the deployment of discrete tools toward the orchestrated interaction of multiple interconnected components within a complex, adaptive system. Classical systems theory outlines an "input-integration-output-feedback" paradigm as a foundational framework for understanding complex system operations 25,26 . This paradigm describes a process whereby external resources are transformed into effective outputs through internal structural processing. Its high level of abstraction provides a macro-level, process-oriented lens for analyzing the intricate ecosystems of the digital age. Digital ecosystems themselves have conceptual origins in the interdisciplinary adaptation and reimagining of the traditional "ecosystem" concept. Their theoretical roots can be traced to Moore's (1993) "business ecosystem" theory, which emphasizes the co-evolutionary capacity of interconnected organizations within an economic community. With the advent of the digital era, a distinct theoretical perspective known as the "digital ecosystem" has emerged. In China, this concept was formally elevated to a policy level in 2021 when the State Council's "14th Five-Year Plan and the Long-Range Objectives Through 2035" called for establishing a digital rules system to foster an open, healthy, and secure digital ecology. Qiu et al. (2025) further proposed that the core dimensions of digital ecosystem theory encompass four interrelated aspects: interconnected elements, actors, interdependent interactions, and developmental evolution. They characterize digital ecosystems as complex systems that maintain relative independence while interacting with natural and social ecosystems. In light of this, this paper develops a systematic "technology input" framework, which is realized through the introduction of emerging digital technologies such as artificial intelligence, serving as the initial driver for public service transformation 27 . However, the realization of technological efficacy depends critically on the "system integration" phase. This phase entails complex coordination and configurational interactions across the five dimensions of the digital ecosystem: digital infrastructure, digital government, digital economy, digital society, and digital capabilities. Technological elements achieve value transformation through deep coupling with the diverse actors, data resources, institutional rules, and application scenarios embedded within these dimensions. This process fundamentally reshapes the underlying logic of public service delivery 28 , transforming it from a traditional supply-oriented model into value-added outcomes that are co-created by multiple stakeholders—including governments, enterprises, and the public—within a digital environment 29 . At the core of this model is artificial intelligence as a technological input. Its potential must be effectively activated and regulated through differentiated configurations of the digital ecosystem as the system integrator, ultimately converting technological potential into tangible "value outputs." These outputs are manifested as structural optimizations of public services across five dimensions: innovation, green development, openness, sharing, and coordination. The theoretical framework is illustrated in Figure 1. 2 Theoretical Mechanisms and Research Pathways 2.1 Digital Ecosystem and Public Service Structural Optimization: A Configuration Perspective In the current era of global digital transformation, building a robust digital ecosystem has become the key driver behind the iterative upgrading of public service structures. However, the digital ecosystem is not the result of a single isolated factor, but rather a complex system comprising five interrelated dimensions: digital infrastructure, digital economy, digital government, digital society, and digital capability. These dimensions are deeply intertwined and mutually reinforcing. Through diverse configurations, they act individually or collectively across different value dimensions of public service structural optimization, jointly shaping the mode, quality, and efficiency of public service delivery and ultimately enabling the co-creation of public value. (1) Digital infrastructure, as the foundational backbone supporting public service development, comprises robust network connectivity, digital hubs, and computing power—collectively forming the underlying architecture for efficient public service operations 30 . High-speed networks facilitate the seamless flow of data and enable cross-domain collaboration, while the intensive use of cloud computing significantly enhances resource utilization efficiency and supports the green and low-carbon transformation of public services. Notably, the rapid advancement of computing systems provides core technical support for complex public service applications such as smart cities. Therefore, building advanced, inclusive, and secure digital infrastructure is an indispensable prerequisite for enhancing the quality of public service delivery. (2) As the most dynamic component, the digital economy serves as a powerful external driver and source of innovation for optimizing public service structures. Emerging technologies and business models—including big data, artificial intelligence, and platform economies—are increasingly integrated into public services, giving rise to innovative service paradigms such as smart healthcare and personalized education 31 . The principles of the sharing economy not only optimize social resource allocation and improve overall operational efficiency but also reflect the core values of green development. Leading technology enterprises, through public-private partnerships, open their mature core capabilities—such as payment and identity authentication systems—to government agencies, significantly improving service convenience and diversity. Moreover, the vigorous growth of the digital economy compels governments to accelerate data openness, unlocking the social value of public data and fostering a virtuous cycle of "building through utilization" in public service optimization. (3) Digital government construction functions as the core engine for public service structural optimization, fundamentally reshaping service delivery models through deep institutional transformation. Institutionally, digital government promotes frameworks such as "one-stop online services" and "integrated online governance," effectively addressing service fragmentation by enabling cross-departmental and cross-level coordination of business processes and data sharing. Simultaneously, digital government actively releases public data to society via established open platforms 32 , safeguarding citizens’ rights to information and oversight while stimulating social innovation. The comprehensive digitization of government services overcomes traditional geographical constraints, significantly advancing the equalization and universal access to basic public services. Overall, the depth and breadth of digital government development directly determine the extent to which public service efficiency can be enhanced. (4) The maturity of a digital society directly reflects citizens’ digital living standards and participatory capacities, serving as a critical factor in determining whether public service reforms achieve broad societal acceptance and effective implementation. In a highly developed digital society, individuals skillfully employ digital tools for expression, oversight, and interaction. This widespread civic engagement shifts public service provision from a "government-led" model toward a "people-centered" approach, enabling more responsive alignment with genuine public needs. Public wisdom and creativity emerge as vital sources of innovation, exemplified by crowdsourced models in environmental monitoring and community governance. A robust digital social ecosystem further strengthens mutual trust between government and citizens, contributing to broader societal cohesion. (5) Digital capability centers on the human element, functioning as the soft power and fundamental enabler for realizing public service structural optimization. It encompasses both citizens’ digital literacy and civil servants’ digital competencies 33 . On one hand, improving nationwide digital literacy is essential for bridging the digital divide and ensuring equitable access to the benefits of digital development across all population groups; otherwise, technological progress may deepen existing inequalities. On the other hand, civil servants equipped with digital thinking and execution skills serve as internal catalysts, empowering governments to proactively adopt new technologies for process reengineering and service innovation. Strong digital capabilities also facilitate effective communication and collaboration among diverse governance actors, promoting coordinated and integrated service delivery. Thus, digital capacity building constitutes the decisive factor in activating the entire digital ecosystem, ensuring that public service reforms remain people-centered and sustainable. In summary, the five dimensions of the digital ecosystem do not operate in isolation. Rather, they interact through interconnected and combinatory mechanisms, synergistically driving the optimization of public service structures. This paper examines the integrated pathways through which the digital ecosystem enhances public service structural reform from a configurational perspective, laying the theoretical foundation for further exploration of the relationships and influence mechanisms among artificial intelligence, the digital ecosystem, and public service structural optimization. 2.2 Complex Mediation Effect: From Configuration Pathways to Mechanism Testing 2.2.1 Direct Effects As a critical technological input, artificial intelligence provides essential enabling capabilities for optimizing public service structures through its advanced computing power and sophisticated algorithms. First, at the innovation level, AI drives innovation in service models and operational formats, enabling predictive regulation by analyzing large-scale risk data to issue early warnings and implement timely interventions prior to safety incidents. It also supports personalized public services, reflecting a transformative shift from "standardization" to "diversification"—a set of revolutionary advancements enabled by AI 34 . Second, in terms of sustainability, AI’s contributions extend well beyond resource optimization. Through millisecond-level monitoring, prediction, and simulation of urban energy systems, building energy consumption, and traffic flows, AI enables precise supply-demand matching and dynamic scheduling while significantly reducing wasteful energy use. For instance, AI-powered smart grids balance fluctuations in renewable energy generation, while intelligent transportation systems effectively minimize emissions caused by traffic congestion, making the operation of public service systems themselves exemplars of green, low-carbon efficiency. Third, at the openness level, AI enhances government transparency and facilitates public participation. By leveraging data mining and visualization tools, AI transforms raw, complex datasets into intuitive charts, actionable insights, and interactive applications. This substantially lowers the barriers for citizens to understand and utilize open government data, thereby enhancing its practical utility and fostering greater civic engagement. Fourth, at the sharing level, AI can intelligently analyze massive, multi-source datasets to more accurately forecast trends, identify risks, and optimize resource allocation. This enables public service providers to precisely identify the needs of diverse regions and population groups, directing resources toward grassroots communities, rural areas, and vulnerable populations while minimizing misallocation and waste. For example, AI can predict changing demands for public transportation, elderly care facilities, and logistics networks, guiding public investment with greater spatial precision. This improves the equitable distribution of urban-rural infrastructure and services, promoting coordinated development from the planning stage onward. In summary, this paper proposes Hypothesis 1: Hypothesis H1: Artificial intelligence has a significant positive impact on the optimization of public service structures. 2.2.2 Indirect Effects The key to AI serving the public sphere lies in the rational construction of digital ecosystems. As a vital technological input, AI’s functionality depends on the systematic integration of digital services and technological ecosystems to achieve the ultimate goal of delivering high-quality public services. This implies that digital ecosystems serve as an essential mediating mechanism between AI and public service structural optimization, manifested specifically in three critical challenges: First, algorithmic dependency and inclusivity challenges. The effective operation of AI relies on high-quality, large-scale data. If digital infrastructure coverage is uneven or public digital literacy is inadequate 35,36 , AI may exacerbate the "digital divide," thereby undermining equitable access to public services. Second, algorithmic bias and fairness challenges. AI systems can replicate and amplify existing societal biases during automated decision-making, raising serious ethical concerns that threaten equity in public service delivery 37 . Third, black-box decision-making and accountability challenges. The opacity of AI-driven decision processes poses significant difficulties for traditional government accountability mechanisms, creating a tension between data-intensive governance and bureaucratic accountability structures. In summary, without robust ecosystem support, AI applications risk generating new crises in fairness, ethics, and governance—even as they enhance operational efficiency. Therefore, adopting a systemic perspective through the lens of the digital ecosystem is essential to mitigate risks associated with AI applications and ensure the realization of public value. As an organic whole, the digital ecosystem provides indispensable support for the healthy and sustainable development of AI. Robust digital infrastructure ensures widespread access to AI services and enables energy-efficient computing; strong digital capabilities constitute fundamental prerequisites for bridging the digital divide and facilitating broad AI adoption; effective digital governance establishes institutional mechanisms such as algorithmic auditing and ethical oversight frameworks to systematically address bias and safeguard fairness. A vibrant digital economy and a mature digital society collectively provide AI with a continuous source of innovation and broad-based societal acceptance. Thus, the digital ecosystem is not passive but functions as a dynamic intermediary—capable of being activated by AI while simultaneously regulating and guiding its deployment. In summary, this paper proposes Hypothesis 2: Hypothesis H2: The digital ecosystem plays a pivotal intermediary role in AI-driven optimization of public service structures, where varying configurations of its components lead to distinct pathways and outcomes. 2.2.3 Research Framework and Core Questions Given this, this paper develops a complex intermediary model from the perspective of the digital ecosystem, examining how AI technology interacts with the digital ecosystem to influence public service structural optimization. It seeks to address the following research questions: How does AI directly impact public service structural optimization? What configurations of the digital ecosystem can effectively drive this transformation? Furthermore, as a complex intermediary, how do different combinations of its internal elements differentially shape the translation of technological inputs into high-quality public service outputs? The complex intermediary model proposed in this paper, illustrated in Figure 2, clearly delineates the complete causal pathway from technological inputs to value outcomes. Research Design 3.1 Complex Mediating Model Confronted with multiple concurrent causal relationships in research on public service structural optimization, traditional single statistical methods struggle to uncover deep-seated interdependencies among variables. Therefore, drawing on the approach of Du et al. (2024), this study integrates dynamic qualitative comparative analysis (QCA) with regression analysis to construct a mixed-methods testing framework 38 . This framework aims to identify multiple equivalent configuration pathways through which the digital ecosystem influences public service structural optimization, while simultaneously assessing the indirect effects of artificial intelligence on public service structural optimization via these mediating configurations. The complex mediation model comprises four steps (see Figure 3): First, examine the M-Y causal chain. Employ dynamic QCA analysis to identify multiple condition combinations influencing the outcome variable—specifically, how different configurations of the digital ecosystem affect the optimization of public service structures. Second, assign values to configuration M, converting dynamic QCA results into mediating variables for regression analysis. Third, test the X-M causal chain to examine the overall effect of AI on the formation of the digital ecosystem (entropy values across five dimensions), i.e., the effect of path a. Step 4: Test the X-Y causal chain to further explore AI's indirect influence through digital ecosystem configurations. Key notes: ① During configuration membership calculation, a case's membership in configuration M is determined by taking the minimum membership value across all condition sets; ② When multiple configuration elements are interdependent, separate regression analyses treat different configurations as independent mediating variables to avoid multicollinearity. 3.2 Data Sources This study utilizes a panel of 30 Chinese provinces (excluding Tibet) covering the period from2019 to 2023 to examine the complex interplay among artificial intelligence, the digital ecosystem, and the structural optimization of public services. Data on AI and the digital ecosystem are drawn from the Digital Ecology Index. Additional variables are sourced from a comprehensive set of official publications, including the China Statistical Yearbook, Health Statistical Yearbook, China Local Digital Service Capability, China Fiscal Yearbook, China Government Transparency Report, China Environmental Statistical Yearbook, China Industrial Statistical Yearbook, and the China City Statistical Yearbook, supplemented by provincial statistical yearbooks and data from the National Bureau of Statistics. Missing values in the dataset were handled using linear interpolation and mean imputation methods to ensure completeness. 3.3 Data and Measurement 3.3.1 Dependent Variable: Artificial Intelligence Given that artificial intelligence is a multidimensional, pervasive general-purpose technology, it is challenging to comprehensively measure its overall development level and application potential at the provincial level using a single absolute indicator. Therefore, this study adopts the "Digital Innovation" sub-index from the 2019-2023 Digital Ecosystem Index, jointly released by the National Engineering Laboratory for Big Data Analysis and Application Technology at Peking University and multiple institutions, as a metric for assessing AI development. 3.3.2 Explanatory Variable: Optimization of public service structure Public service structure optimization is defined as advancing the public service system from "ensuring basic coverage and broad accessibility" toward "enhancing quality, expanding capacity, and achieving balanced accessibility," with the fundamental goal of meeting the people's growing needs for a better life. This ultimately aims to realize a new stage of development characterized by higher levels, higher quality, greater efficiency, greater fairness, and greater sustainability. Therefore, drawing on Zhang (2025) 39 , this paper constructs an indicator system for optimizing public service structures based on development levels across five dimensions: innovation, green development, openness, sharing, and coordination (see Table 1). Table 1: Indicator System for Public Service Structure Optimization Primary Indicators Second-Level Indicators Measurement Method Attribute Public Service Innovation Development Level Healthcare Talent Service Level Number of licensed physicians, registered nurses, and pharmacists / Total population + Digital Education Service Level Number of terminals in public library e-reading rooms + Public Service Green Development Level Green coverage Green coverage rate in built-up areas + Environmental Protection Investment Environmental Protection Expenditures / General Public Budget Expenditures + Waste treatment level Public Household Waste Harmless Treatment Rate + Public Service Open Development Level Government Service Openness Level Government Transparency Index + Level of Trade in Services Openness Level of Trade in Services Openness + Public Service Shared Development Level Level of Shared Healthcare Services Number of Beds in Public Healthcare Institutions / Total Population + Level of Shared Educational Services Public library collection size per total population + Public Service Coordinated Development Level Level of Coordination in Urban and Rural Elderly Care Services Number of Urban Pension Insurance Participants / Number of Rural Pension Insurance Participants - Level of Coordination in Urban and Rural Transportation and Communications Per capita transportation and communication expenditure of urban residents / Rural residents' transportation and communication expenditure - Urban-Rural Per Capita Healthcare Coordination Level Per capita healthcare expenditure of urban residents / Per capita healthcare expenditure of rural residents - 3.3.3 Mediating Variable: Digital Ecosystem Since the Digital Ecosystem Index report aggregates underlying indicators into standardized indices through its proprietary weighting system, this study directly adopts the published final sub-index values. Based on the Digital Ecosystem Index framework, the digital ecosystem is categorized into five dimensions: ① Digital Infrastructure, encompassing infrastructure development, data resources, and policy environment; ② Digital Government, including the Online Government Service Capability Index, Smart Environmental Protection Index, Digital Government Development Index, and Rural Digital Governance Index; ③ Digital Economy, comprising the Big Data Industry Development Index, Artificial Intelligence Industry Development Index, Digital Industry Electricity Consumption Index, Digital Economy Investor Confidence Index, Enterprise Digital Transformation Index, SME Digitalization Index, Micro and Small Enterprise Digital Development Index, and Rural Digital Economy Index; ④ Digital Society, covering the Digital Inclusive Finance Index, Digital Lifestyle Index, Social Dispute Search Index, Digital Convenience Payment Index, and Rural Digital Society Index; ⑤ Digital Capability, consisting of Digital Talent and Digital Security. These five dimensions collectively constitute an integrated system that comprehensively reflects the overall state of a region’s digital ecosystem. 3.3.4 Control Variables To prevent omitted variable bias from affecting estimation results, a set of control variables was incorporated based on established theoretical and empirical research 40,41 . ① Government intervention, measured by the ratio of local public budget expenditure to GDP; ② Urbanization rate, calculated as the proportion of urban population to total population; ③ Industrial structure, captured by the share of tertiary industry value-added in GDP; ④ Level of financial development, measured by the sum of deposit and loan balances of financial institutions as a percentage of GDP. Empirical Analysis Results 4.1 Configural Analysis of Digital Ecosystems Optimizing Public Service Structure To ensure consistency in calibration standards over time and enhance the longitudinal comparability of results, this study employs a theory-driven dynamic QCA fixed-anchor calibration approach 42 , setting threshold values for full membership, cross-case intersections, and full non-membership at the 75th percentile, 50th percentile, and 25th percentile, respectively. 4.1.1 Necessity Condition Analysis Prior to conducting the configurational analysis, the necessity of each antecedent condition must be verified. When the consistency level is greater than or equal to 0.9, the condition variable can be considered a necessary condition for the outcome variable 43 . However, in panel data QCA analysis, consistency-adjusted distance must be further employed to examine whether necessary conditions exhibit temporal or case effects, ensuring the reliability of aggregated consistency. Calculated via formula, a consistency-adjusted distance below 0.2 indicates higher precision in aggregated consistency and coverage 44 . When the adjusted distance exceeds 0.2, researchers should further investigate its necessity 45 . Table 2 shows that the consistency levels for all conditions are below 0.9, indicating no necessary conditions affecting the optimization of high public service structures. To verify the stability of these results, the consistency distance was adjusted. Among the groups—non-high digital government, high digital society, non-high digital society, non-high digital economy, and non-high digital capability—the adjusted distances exceeded 0.2. Annual consistency level checks revealed excessively large adjusted distances between variable groups. When consistency levels ranged between 0.1 and 0.9 and inter-group coverage exceeded 0.5, these conditions were deemed non-essential. Table 2 Necessity Analysis Conditions Condition variable High-Level Public Service Structural Optimization Low-Level Public Service Structural Optimization Aggregate Consistency Aggregate Coverage Intergroup Consistency Adjustment Distance Intra-Group Consistency-Adjusted Distance Aggregate Consistency Aggregate Coverage Inter-Group Consistency-Adjusted Distance Intra-Group Consistency-Adjusted Distance High-digit foundation 0.808 0.787 0.081 0.460 0.348 0.343 0.258 0.690 Non-high digit foundation 0.326 0.331 0.200 0.719 0.784 0.805 0.151 0.311 High Digital Government 0.788 0.784 0.148 0.449 0.338 0.340 0.510 0.702 Non-high digital government 0.337 0.335 0.391 0.684 0.785 0.789 0.223 0.276 High Digital Society 0.759 0.756 0.246 0.460 0.359 0.362 0.617 0.638 Non-high digital society 0.359 0.357 0.527 0.661 0.758 0.761 0.322 0.293 High Digital Economy 0.842 0.829 0.171 0.518 0.311 0.310 0.432 0.748 Non-high digital economy 0.299 0.300 0.499 0.730 0.828 0.842 0.206 0.270 High Numerical Ability 0.839 0.832 0.104 0.420 0.309 0.310 0.304 0.742 Non-High Numerical Ability 0.304 0.303 0.252 0.736 0.833 0.840 0.090 0.334 4.1.2 Configurational Analysis Configuration analysis is the distinctive core function of the QCA method, primarily used to reveal the impact of different condition combinations on the outcome variable. Referencing existing research and considering practical circumstances, this study ultimately selected a consistency threshold of 0.85, a case frequency of 2, and a PRI threshold of 0.7. It prioritized intermediate solutions while supplementing with minimal solutions to identify core and peripheral conditions. Table 3 presents the overall configuration analysis results. The aggregated consistency of the high-public-service-structure optimization solution reached 0.916, exceeding the 0.75 threshold. Moreover, the inter-group consistency adjustment distance for each configuration remained below 0.2, while the highest intra-group consistency adjustment distance was 0.212. This indicates strong explanatory power for the aggregated consistency, confirming that these three configurations constitute sufficient conditions for achieving high public service structure optimization. This outcome comprises three configurations: namely Ecologically Mature Type (H1), Administratively Driven Type (H2), and Socially Empowered Type (H3). Subsequently, each configuration undergoes detailed analysis and comparative evaluation. (1) Mature Ecosystem Model (H1). This path relies on digital infrastructure, digital government, digital society, and digital economy as core prerequisites, with its intrinsic momentum stemming from deep inter-system coordination and virtuous cycles. Typical examples of this path include developed provinces and cities like Beijing, Shanghai, Jiangsu, Zhejiang, and Guangzhou. These regions lead China across multiple dimensions—hardware construction, institutional guidance, industrial vitality, and social applications—forming a development pattern characterized by comprehensive leadership and balanced elements. Within this system, while citizens' digital capabilities are not explicitly highlighted as a primary condition, their overall high level of development has long become the foundational bedrock for societal operations. The formation of this pathway stems from three key factors: First, top-tier digital infrastructure provides a robust foundation for digitalization across all sectors; second, forward-looking digital governance frameworks and efficient institutional supply offer rule-based guidance and policy safeguards for the coordinated evolution of the entire system; third, vibrant digital economic innovation and widespread digital societal applications have not only spawned new business models and formats but also compelled service optimization and upgrading. Thus，the ecosystem maturity path represents a more advanced developmental form emphasizing systemic synergy，innovation-driven growth，and value-led guidance，serving as a benchmark和model for regional digitalization. Within such an ecosystem，public services continuously enhance quality和efficiency，ultimately maximizing public value realization. (2) Administration-Driven Model (H2). This path centers on digital infrastructure, digital government, digital economy, and digital capabilities as core prerequisites, representing a government-led, efficiency-first development model. While its case distribution overlaps somewhat with the Ecologically Mature Model (H1), its most representative provinces are municipalities like Tianjin and Chongqing. This indicates that even in developed regions with relatively mature digital ecosystems, strong government guidance remains a crucial driver for upgrading public service quality. Leveraging the centralized and efficient advantages of their administrative systems, these regions achieve rapid progress in key areas like government digitalization through top-down design, resource allocation, and targeted policies—even when the broader social digital application environment is still developing. The effectiveness of this approach hinges on three pillars: first, unified institutional arrangements provide stable expectations and directional guidance for development; second, efficient execution systems and performance incentives ensure the thorough implementation of policy objectives; third, the government's control over critical resources and infrastructure lays the foundation for coordinated action across the entire region. Consequently，the administratively driven model significantly shortens innovation diffusion cycles in areas requiring rapid response and concentrated resource allocation，achieving a "point-to-area" development effect. (3) Social Empowerment Pathway (H3). This pathway centers on digital infrastructure, digital society, digital capabilities, and digital government as core prerequisites. Unlike traditional economic development models that prioritize growth first, this approach embodies an inclusive and universally accessible path of social empowerment. Examining specific cases, representative provinces for this path include Hebei and Henan. In these regions, the contribution of digital economy industries remains relatively weak, and government efficiency is not high. However, as key areas actively advancing the optimization of public service structures, the socially empowered development path can achieve this goal by strengthening digital infrastructure construction, promoting the widespread application of digital society, and enhancing the digital capabilities of the entire society. The reasons for this path emerging are as follows: first, robust digital infrastructure provides a solid foundation for the development of public services; second, effective digital government design and policy implementation provide institutional safeguards for the digital transformation of public services and the enhancement of public digital literacy; third, widespread digital societal applications and public demand create a synergistic force for the reapplication of digital technologies，thereby compelling improvements in public service quality. Thus，the socially empowered development path represents a model that places greater emphasis on social participation，demand-driven growth，and inclusive sharing. In summary, Pathways H1, H2, and H3 are not mutually exclusive but exhibit a progressive, convergent, and diffuse evolutionary trend. On one hand, in terms of driving logic, Pathway H1 represents a mature ecosystem that relies on the comprehensive synergy of the five key elements of the digital ecosystem to form a self-sustaining virtuous cycle within the system, embodying a comprehensive, naturally evolved advanced form; Pathway H2 highlights the government's leading role, concentrating resources on breakthroughs in infrastructure, industrial cultivation, and capacity building for key groups. It leverages administrative power to compensate for deficiencies in the societal digital ecosystem, manifesting as a top-down, efficiency-first intervention model; Pathway H3 emphasizes leveraging digital foundations as a base to drive inclusive public service coverage under relatively weak digital economic momentum. It harnesses dual forces of government guidance and social participation, presenting an inclusive approach characterized by foundational prioritization and demand-driven growth. Trend-wise，Pathway H1 signifies progress from high-level individual elements toward highly efficient synergy. Pathway H2 signifies a gradual shift in the government's role from builder to regulator and enabler，focusing on rule-setting and platform-building. Pathway H3 enhances the socialization of public services by stimulating digital vitality at the societal level，thereby driving the overall upgrade of digital government and digital industries. Table 3 Configuration Analysis Results Condition Variable Configuration Analysis - High Ecosystem Maturity Type Administrative-Driven Type Socially Empowered Type Digital Infrastructure ● ● ● Digital Government ● ● ● Digital Economy ● ● Digital Society ● ● Digital Competence ● ● Consistency 0.931 0.956 0.945 PRI 0.915 0.946 0.931 Coverage 0.616 0.606 0.666 Unique Coverage 0.021 0.01 0.07 Intergroup Consistency Adjustment Distance 0.174 0.139 0.148 Intragroup Consistency Adjustment Distance 0.212 0.121 0.092 Overall Consistency 0.916 Overall PRI 0.896 Overall Coverage 0.697 Note:● indicates the presence of a core condition. All conditions shown are core conditions for their respective configurations. 4.1.3 Robustness Test Robustness tests were conducted on the digital ecosystem configurations yielding high public service structural optimization: ① Increasing the consistency threshold from 0.85 to 0.9; ② Raising the case frequency from 2 to 3; ③ Elevating the PRI from 0.7 to 0.75. The two configurations generated through these three approaches were essentially consistent with the two solutions in the existing configurations. Therefore, the results can be considered robust. 4.2 Analysis of the Direct Effects of Artificial Intelligence on Public Service Structural Optimization 4.2.1 Descriptive Analysis The mediating variables in this section are the three digital ecosystem configurations that yielded high public service structure optimization in the sufficiency analysis. By calculating the membership degree of each configuration group, new mediating variable values were derived and subjected to independent regression analysis. The specific measurement methods and descriptive analysis for each variable are presented in Table 4. Table 4 Descriptive Statistics Variable Type Variable Name Measurement Description Observed Value Mean Standard Deviation Minimum Maximum Independent Variable Artificial Intelligence Digital Patent Index 150 0.368 0.087 0.196 0.65 Mediating Variable Configuration 1 Set membership degree of each city in the corresponding digital ecosystem configuration 150 0.329 0.385 0 0.999 Configuration 2 150 0.315 0.377 0 0.999 Configuration 3 150 0.35 0.386 0 0.999 Dependent Variable Public Service Structure Optimization Composite Index for Public Service Structure Optimization 150 0.301 0.255 0 1 Control Variables Government Involvement Local Public Budget Expenditure to GDP Ratio 150 0.247 0.101 0.107 0.643 Urbanization Rate Proportion of urban population relative to total population 150 0.653 0.096 0.49 0.89 Industrial Structure Upgrading Value Added of the Tertiary Industry as a Percentage of GDP 150 1.584 0.82 0.751 5.69 Financial Level Total Deposits and Loans of Financial Institutions 150 3.785 1.088 2.38 8.16 4.2.2 Regression Analysis The regression results are shown in Table 5. Column (1) and Column (2) present the regression results for AI's impact on optimizing public service structures with and without the inclusion of control variables, respectively. Analysis indicates that regardless of whether control variables are included, the coefficients remain significantly positive, supporting the validity of Hypothesis H1. Given that subsequent regression analysis employs configuration membership values as the mediating variable, the initial regression on raw digital ecosystem data aims to examine whether AI development holistically promotes the digital ecosystem at a macro level. Columns (3) and (4) present regression results for AI's impact on the digital ecosystem without and with control variables, respectively. Analysis indicates that AI exerts a significant positive influence on the digital ecosystem. Table 5 Regression Analysis Results Variable （1） （2） （3） （4） Artificial Intelligence 0.283*** (-17.71) 0.146*** (-7.4) 0.692*** (-23.98) 0.392*** （-11.34） Government Involvement -0.219*** (-4.53) -0.832*** (-9.85) Urbanization Level 0.180** （-3.27） 0.0211 （-0.22) Industrial Structure Upgrading -0.0148** (-2.76) 0.0238* （-2.54） Financial Development 0.0319*** (-5.5) 0.0311** (-3.07) Fixed Effects Yes Yes Yes Yes Constant Term 0.283*** （-45.14） 0.164*** (-5.38) 0.270*** (-23.88) 0.397*** (-7.45) R2 0.6806 0.8199 0.7949 0.8913 Observations (N) 150 150 150 150 Note：* p<0.05， ** p<0.01，*** p<0.001. 4.2.3 Robustness Tests To ensure the robustness of regression results, robustness tests were conducted in four aspects: First, lagging by one period. Given that AI technology applications require time validation, the explanatory variables were lagged by one period and regressed again. Results are shown in Column (1) of Table 6. Second, replacing explanatory variables. Regression was performed using AI word frequency, number of AI enterprises, number of AI industry employees, and industry average salary entropy. Results are shown in Column (2) of Table 6. Third, tail trimming. To mitigate the interference of extreme values on regression outcomes, core variables underwent two-tailed trimming at the 1% level. Results are presented in Column (3) of Table 6. Fourth, we added control variables. We conducted regression analyses incorporating foreign investment levels and total factor productivity, with results shown in Column (4) of Table 6. Fifth, we implemented two-way fixed effects，further controlling for provincial fixed effects in the regression. The above test results indicate that all coefficients for artificial intelligence are positively significant，confirming the robust validity of the regression results. Table 6 Robustness Test Results Variable （1） （2） （3） （4） （5） Artificial Intelligence 0.157*** （-8.33） 0.277*** （-8.01） 0.147*** （-7.52） 0.122*** （-5.56） 0.0375* （-2.12） Control Variables Controlled Controlled Controlled Controlled Controlled Fixed Effects Yes Yes Yes Yes Yes Constant Term 0.141*** (-4.62) 0.182*** （-6.14） 0.165*** （-5.49） 0.150*** （-1.47） -0.0293 (-0.16) R2 0.8580 0.8283 0.8224 0.8254 0.9800 Observations (N) 120 150 150 150 150 4.3 Analysis of Complex Mediation Mechanisms 4.3.1 Mediating Effects This study further analyzes the impacts and mechanisms of three digital ecosystem configurations on the optimization of public service structures through artificial intelligence, based on the mediation model, to identify indirect effects. The results are shown in Table 7. (1) Mechanism of the Mature Ecosystem Application. As shown in columns (1) and (2) of Table 7, after introducing the path of the mature ecosystem (Configuration 1), the coefficient of AI decreased from 0.146 to 0.107, while the coefficient of M1 was 0.0494, passing the 1% significance level test. Specifically, the synergistic development of four key elements—digital infrastructure, digital government, digital society, and digital economy—forms a virtuous cycle within the digital ecosystem. AI leverages advanced infrastructure to enable data flow, promotes cross-departmental coordination through high-level government mechanisms, and gains sustained momentum from an active digital economy. This pathway not only reduces friction in AI implementation but also continuously enhances the quality and efficiency of public services through internal systemic feedback loops. (2) Administration-Driven Application Mechanism. As shown in columns (1) and (3) of Table 7, the coefficient for the administration-driven pathway (Configuration 2) reached 0.0561 and was statistically significant at the 1% level. After introducing M2, the direct effect of AI further decreased to 0.103. The core advantage of this pathway lies in government-led precision efforts, indicating that even in environments with relatively weak digitalization or industrial foundations, the government can rapidly overcome constraints in key areas through robust policy guidance, resource allocation, and performance incentives. This approach clears institutional barriers for AI adoption and creates favorable implementation conditions. (3) Social Empowerment Application Mechanism. As shown in columns (1) and (4) of Table 7, the coefficient for the social empowerment pathway (Configuration 3) is 0.0656, while the direct effect of AI decreases to 0.0864. This indicates that in regions with relatively limited contributions to the digital economy but active social participation, AI can still leverage inclusive digital infrastructure, widespread social applications, and public digital literacy to form a virtuous cycle of "demand-driven service-responsive." This pathway does not rely on high-intensity economic investment but instead adopts a socially demand-oriented approach. By enhancing accessibility, inclusiveness, and participation in public services, it achieves more sustainable structural optimization, offering a replicable path for developing regions. Overall, the three pathways exhibit distinct complementarity: the ecologically mature model represents holistic synergy under optimal system conditions; the administratively driven model embodies targeted breakthroughs under government leadership; while the socially empowered model demonstrates inclusive development propelled by societal forces. Together, they constitute a multifaceted implementation mechanism for AI-empowered public services, offering practically viable pathway choices for regions at different developmental stages and with varying resource endowments. Therefore, we accept H2. Table 7 Complex Mediating Outcomes Variable （1） （2） （3） （4） Configuration 1 0.0494*** (-3.79) Configuration 2 0.0561*** (-4.1) Configuration 3 0.0656*** (-4.78) Artificial Intelligence 0.146*** (-7.4) 0.107*** (-5) 0.103*** (-4.79) 0.0864*** (-3.88) Control Variables Controlled Controlled Controlled Controlled Fixed Effects Yes Yes Yes Yes Constant Term 0.164*** (-5.38) 0.172*** (-5.91) 0.178*** (-6.13) 0.155*** (-5.46) R2 0.8199 0.8357 0.8382 0.8442 Observations (N) 150 150 150 150 4.3.2 Bootstrap Robustness Test This study employs Bootstrap resampling with 5000 iterations for testing. As shown in Table 8, the confidence intervals for the mediating effects across all three configurations do not include zero, indicating statistically significant results. This confirms that the digital ecosystem configuration exerts a stable mediating effect in the process of AI influencing the optimization of public service structures. Specifically, the strength of the mediating effects across the three pathways shows distinct differences. First, the mediating effect accounts for 26.5% in Configuration 1 (Ecosystem Maturity Type), indicating that approximately one-quarter of AI's impact on optimizing public service structures is achieved through a highly synergistic digital ecosystem. Only when all elements of the digital ecosystem reach a certain level of maturity can its systemic synergistic effects be fully unleashed. Second, the mediating effect in Configuration 2 (Administrative-Driven Type) is stronger, accounting for 29.6%. This indicates that at the current stage of development, the government's leading role through top-level design, resource allocation, and policy constraints serves as a critical transmission pathway in the AI-driven optimization of public service structures. Compared to the ecosystem maturity model, this pathway does not require waiting for all elements to mature naturally. Instead, it leverages administrative power for selective breakthroughs in key areas, thereby demonstrating higher efficiency in accelerating the rapid implementation of AI applications. Finally, Configuration 3 (Social Empowerment Type) exhibits the most pronounced mediating effect, accounting for 40.9%. This result confirms that broad societal participation and enhanced public digital literacy serve as a critical lever. When citizens possess high digital capabilities and strong engagement awareness, they generate powerful demand pull and oversight functions. This drives public service departments to improve quality and efficiency while simultaneously providing rich application scenarios and data foundations for AI deployment. Furthermore, the study reaffirms that AI's direct effect on optimizing public service structures remains significant, though its coefficient exhibits a decreasing trend: from 0.1075 in Configuration 1 to 0.0864 in Configuration 3. This shift indicates that as the mediating mechanism within the digital ecosystem strengthens, AI's role in optimizing public service structures is increasingly realized through positive feedback loops within the system, thereby reducing the pressure for direct intervention. In summary, a mature, collaborative, or active digital ecosystem can more effectively transmit and amplify the technological dividends of AI. Table 8 Bootstrap Robustness Test Mediating Variable Effect Type Coefficient Standard Deviation 95% Confidence Interval Mediation Proportion Configuration 1 Indirect 0.0387762 0..0131539 0.012995 0.0645574 26.5% Direct 0.1074764 0.0211699 0.0659841 0.1489687 Configuration 2 Indirect 0.0432523 0.0135792 0.0166375 0.0698671 29.6% Direct 0.1030003 0.0212285 0.0613931 0.1446074 Configuration 3 Indirect 0.0598909 0.0174463 0.0256968 0.094085 40.9% Direct 0.0863617 0.02304 0.041204 0.1315193 Conclusions and Countermeasures 5.1 Research Conclusions This study employs dynamic QCA and regression analysis, examining 30 provinces in China (excluding Tibet) from 2019 to 2023. Integrating digital ecosystem theory and configurational perspectives, it systematically investigates the complex mechanisms through which artificial intelligence drives structural optimization in public services. Notably, it reveals the critical mediating role of digital ecosystems in this process, leading to the following conclusions: (1) The configuration analysis results indicate that the digital ecosystem plays a crucial role as a multifaceted intermediary. This study innovatively employs three pathways derived from dynamic QCA: the ecologically mature type, the administratively driven type, and the socially empowered type. These pathways are not mutually exclusive but instead exhibit a progressive, convergent, and diffusive evolutionary trend toward optimizing public service structures. Specifically, different Chinese provinces can select distinct development paths based on their developmental stage, resource endowments, and core strengths. For instance: Eastern coastal provinces may prioritize the Ecologically Mature Path, leveraging mature subsystems to pursue higher levels of digital ecosystem synergy; Certain municipalities can effectively employ the Administratively Driven Path to achieve leapfrog development in key industries like information technology and biopharmaceuticals; Central and western regions, with their vast social potential, can deepen the social empowerment path to leverage public strengths and social capital, charting a "people-centered" route for structural optimization. (2) The findings from the complex intermediary mechanism test reveal dual pathways for AI-enabled structural optimization of public services: on one hand, AI promotes the optimization of public service structures; on the other, it drives such optimization at different levels through three distinct types of digital ecosystem intermediaries—ecologically mature, administratively driven, and socially empowered. This underscores that advancing public service modernization requires not only prioritizing AI technology R&D and application but also cultivating a complementary digital ecosystem. This fosters a virtuous cycle of technological empowerment and systemic optimization, ultimately maximizing public value. 5.2 Recommendations (1) Implement differentiated and targeted AI development strategies. The three development pathways identified in this study—ecologically mature, administratively driven, and socially empowering—offer diverse strategic options for regions at different stages of development. Regions should avoid homogeneous competition through a "one-size-fits-all" approach. Instead, they should leverage their unique resource endowments and development priorities to select the most suitable entry points and implement differentiated, targeted AI development strategies. First, for developed regions like Beijing and Shanghai with robust digital infrastructure and vibrant markets, efforts should focus on building open, collaborative innovation platforms. This involves facilitating orderly data flow and value realization while encouraging multi-stakeholder participation—including enterprises and research institutions—in scenario-based innovation. The goal is to cultivate a self-evolving, dynamic ecosystem. Second, municipalities like Tianjin and Chongqing, characterized by strong administrative leadership and efficient resource allocation, should concentrate on dismantling "data silos." Through robust top-level design and cross-departmental coordination, they should advance data integration and business process reengineering. Prioritizing breakthroughs in vertical domains such as government services and urban management can drive the entire digital ecosystem's development by significantly enhancing administrative efficiency. Finally, for regions like Henan and Hebei with diverse social needs and active grassroots innovation, policies should moderately tilt downward to encourage community involvement; social organizations; citizen participation in co-governance. By establishing low-barrier tool platforms and incentive mechanisms, AI technology can be leveraged to address pain points in elderly care, education, healthcare, and other public welfare sectors, thereby stimulating the vitality和 resilience of public services from the bottom up. (2) Refine a multi-stakeholder collaborative digital ecosystem governance system. Deep AI application in public services requires joint participation from government, market, and societal forces. Society-wide efforts should actively build a collaborative mechanism integrating government, enterprises, academia, research, and end-users. Guided by government regulations and targeting enterprise digital-intelligence development, proactive exploration of partnerships with universities and research institutions is essential to effectively align digital technology supply with societal demand. Additionally, governments should vigorously advance the centralized and green deployment of computing power alongside the construction of data resource systems, fully leveraging existing social infrastructure resources to avoid redundant investments in funds and projects. Concurrently, regulatory authorities must accelerate the refinement of AI-related laws and standards—particularly establishing clear rules regarding data security, algorithmic transparency, and liability determination—to provide robust safeguards for the dual-engine drive of technological and institutional innovation in public services. (3) Strengthen information risk prevention and control while building a digital value-driven system. While vigorously promoting AI applications, we must maintain a clear awareness of risks and reinforce a human-centered value orientation. First, establish an evaluation system centered on public satisfaction and sense of fulfillment, eliminating formalism driven by "technology for technology's sake." Second, prioritize the inclusivity and accessibility of digital technology applications. Governments should implement measures such as age-friendly renovations, information accessibility initiatives, and preserving traditional service channels to ensure equal and seamless access to intelligent services for all groups, including the elderly and persons with disabilities. Finally, establish dynamic evaluation and corrective mechanisms to ensure AI development and application consistently align with social ethics and public interest, achieving the organic integration of technological empowerment and humanistic care. 5.3 Limitations and Future Directions Although this study identifies diverse pathways for AI-empowered public service structural optimization through configuration analysis and confirms the critical mediating role of the digital ecosystem, certain limitations persist, suggesting avenues for future research. First, the current study focuses exclusively on the complex interaction mechanisms between AI and public service structural optimization at the provincial level from 2019 to 2023. Extending the temporal scope to include data from the past decade would enable a more comprehensive understanding of AI's evolutionary trajectory and long-term impacts. Second, future studies could disaggregate AI technologies—such as generative AI and computer vision—to systematically examine their distinct operational mechanisms and comparative effectiveness across various public service domains, thereby enhancing policy relevance and technological specificity. Declarations A Data Availability Statement： The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. CRediT authorship contribution statement: a Chunlin Xiong: Writing-review and editing, Funding acquisition, Supervision b*Ren Fan: Writing – original draft, Writing – review and editing, Methodology, Software, Validation, Formal analysis c Yi Liu: Investigation Competing interests The author(s) declare no competing interests. Funding: This research was supported by the Key Project of the National Social Science Fund "Research on the Mechanism of Generating Effectiveness in Rural Digital Governance and its Enhancement Path"(24AZZ010);National Social Science Foundation’s Post-funded Project "Evaluation and Optimization of Agricultural and Rural Informatization Policies"(22FGLB006); Hunan Province Social Science Foundation’s “Academic Hunan” High-quality Cultivation Project "Mechanism Innovation for Improving Rural Grassroots Governance Efficiency in the Era of Big Data"(23ZDAJ010); The Key Project of Changsha Soft Science Research Program: Research on Dynamic Evaluation, Simulation Prediction, and Optimi-zation Path of High Quality Development of Rural Digital Economy in Changsha City (KH2502003); Hunan Agricultural University Scientific Research Project "Research on the Driving Mechanism and Optimization Paths of Rural Digital Governance" (25SK015);Hunan Agricultural University Scientific Research Project "Theoretical and Practical Innovations in Digital Rural Development" (25SK027). References Mikhaylov, Slava Jankin, Marc Esteve, and Averill Campion. "Artificial intelligence for the public sector: opportunities and challenges of cross-sector collaboration." Philosophical transactions of the royal society a: mathematical, physical and engineering sciences 376.2128 (2018): 20170357. Latupeirissa, J. J. P., Dewi, N. L. Y., Prayana, I. K. R., Srikandi, M. B., Ramadiansyah, S. A., & Pramana, I. B. G. A. Y. (2024). Transforming public service delivery: A comprehensive review of digitization initiatives. Sustainability, 16(7), 2818. Kuziemski, Maciej, and Gianluca Misuraca. "AI governance in the public sector: Three tales from the frontiers of automated decision-making in democratic settings." Telecommunications policy 44.6 (2020): 101976. Cao, H.J., & Xiong, Z.Q. (2025). Co-creation of value in digital public services: Evolutionary origins, theoretical framework, and operational mechanisms. Theoretical Exploration, (03), 56-66. Kim, Younhee, Seunghwan Myeong, and Michael J. Ahn. "Living labs for AI-enabled public services: functional determinants, user satisfaction, and continued use." Sustainability 15.11 (2023): 8672. Chang, Jiang, Su, Xin, and Shen, Feiwei. "Exploring New Models of Generative AI-Empowered Public Services." Scientific Decision Making, 2025, (10): 231-244. He, Yuanlang, and Yuan, Jianhong. "AI Development and Enhancement of New Quality Productivity: Theoretical Mechanisms and Empirical Tests." Science and Technology Progress and Policy, 2025, 42(11): 1-11. Wirtz, Bernd W., Jan C. Weyerer, and Carolin Geyer. "Artificial intelligence and the public sector—applications and challenges." International Journal of Public Administration 42.7 (2019): 596-615. Mergel, Ines. "Agile innovation management in government: A research agenda." Government Information Quarterly 33.3 (2016): 516-523. Tan Haibo, Ye Wei. Digital Technology Empowering Public Services: Internal Mechanisms and Primary Approaches [J]. Administrative Forum, 2024, 31(02): 103-110. Margetts, Helen, and Cosmina Dorobantu. "Rethink government with AI." Nature 568.7751 (2019): 163-165. Ølnes S, Ubacht J, Janssen M. Blockchain in government: Benefits and implications of distributed ledger technology for information sharing[J]. Government information quarterly, 2017, 34(3): 355-364. Burlacu S, Dobre F, Mateescu V M, et al. From bureaucracy to e-Government: Digital transformation of Public Administration in Romania for a sustainable future[J]. European Journal of Sustainable Development, 2025, 14(3): 289-289. Janowski, Tomasz. "Digital government evolution: From transformation to contextualization." Government information quarterly 32.3 (2015): 221-236. Yao Jiaquan, Zhang Kunpeng, Guo Lipeng, et al. How Does Artificial Intelligence Enhance Corporate Productivity? — A Perspective Based on Labor Skill Structure Adjustment [J].Management World, 2024, 40(02): 101-116+133+117-122. Farrell, Eimear, et al. "Artificial intelligence for the Public Sector." (2023). Garcia Alonso, Roberto, and Sebastian Lippez-De Castro. "Technology helps, people make: A smart city governance framework grounded in deliberative democracy." Smarter as the new urban agenda: A comprehensive view of the 21st century city. Cham: Springer International Publishing, 2015. 333-347. Burlacu S, Dobre F, Mateescu V M, et al. From bureaucracy to e-Government: Digital transformation of Public Administration in Romania for a sustainable future[J]. European Journal of Sustainable Development, 2025, 14(3): 289-289. Liu Haimin, Wu Hailong. "Problem Patterns, Multiple Causes, and Countermeasures of Digital Formalism in Educational Governance." Modern Education Management, 1-14 [2025-10-24]. Linders, Dennis. "From e-government to we-government: Defining a typology for citizen coproduction in the age of social media." Government information quarterly 29.4 (2012): 446-454. Radanliev, Petar, et al. "Ethics and responsible AI deployment." Frontiers in Artificial Intelligence 7 (2024): 1377011. Garcia-Castro, Roberto, and Miguel A. Ariño. "A general approach to panel data set-theoretic research." Journal of Advances in Management Sciences & Information Systems 2.63-76 (2016): 526. Li Deguo and Chen Yiqing. "Risks and Prevention of Artificial Intelligence-Powered Public Services from the Perspective of 'Normal Accidents'." *Social Sciences in Chinese Universities*, 2025, (05): 145-155+159. Li Liqing, Luan Wenting. Generative Artificial Intelligence in Public Service Governance: Mechanisms of Action and Risk Mitigation Strategies. *Journal of Nanchang University (Humanities and Social Sciences)*, 2025, 56(03): 113–124. Hu Hongbin. The Driving System, Operational Mechanism, and Pathways for Enhancing the Development of an Education Nation in the New Era: A Systemic Analytical Framework. Journal of Southwest Minzu University (Humanities and Social Sciences Edition), 2024, 45(07): 186-195. Wu Lingling. On the Systemic Allocation and Optimization Pathways of Library Resources: A Novel Analytical Framework. Library Journal, 2022, (02): 101-106. Busuioc, Madalina. "Accountable artificial intelligence: Holding algorithms to account." Public administration review 81.5 (2021): 825-836. Osborne S P, Cucciniello M, Nasi G, et al. Digital transformation, artificial intelligence and effective public services: challenges and opportunities[J]. Global Public Policy and Governance, 2022, 2(4): 377-380. Wang P. Connecting the parts with the whole: toward an information ecology theory of digital innovation ecosystems[J]. MIS quarterly, 2021, 45(1): 397-422. Xiang Meng, Wang Zhigang. Research on Data Element Market Governance Strategies from a Digital Ecology Perspective[J]. Xinjiang Social Sciences, 2023,(06):54-61. Wang Jiandong, Dou Yue, Ren Junxia, et al. From "Holistic Services" to "Intelligent Entity Services": Research on Innovation in Government Public Service Models Under Digital Transformation [J]. E-Government, 2023, (01): 12-22. Wang Jing, Zhang Hong. Application of Artificial Intelligence in Digital Government Construction and Its Path to Legalization: Analysis Based on DeepSeek's Application Scenarios in Digital Administration [J]. Chinese and Foreign Law, 2025, 37(03): 625-643. Zhang Hongchun, Yang Huan. A Digital Literacy Framework for Civil Servants in the Context of Digital Government: A Conceptual Model [J]. E-Government, 2023, (01): 110-124. Cai Zhenhua, Zhao Youhua. Governance of Public Service Demand in the Era of Artificial Intelligence: Drivers and Directions [J]. Ningxia Social Sciences, 2020, (02): 47-54. Bu Yanjun. Analysis of Power Diffusion Trends and Challenge Responses in the Context of Artificial Intelligence Development [J]. Studies in Science of Science, 2025, 43(03): 608-616. Xue Xinyi, Zhou Jie. Analysis of the Driving Mechanism and Countermeasures for Promoting High-Quality Development of Public Cultural Services by Generative Artificial Intelligence [J]. Library Theory and Practice, 2025, (05): 14-23. Bonezzi, Andrea, and Massimiliano Ostinelli. "Can algorithms legitimize discrimination?." Journal of Experimental Psychology: Applied 27.2 (2021): 447. Du Yunzhou, Sun Ning, Liu Qiuchen. Developing and Analyzing Complex Mediating Models Using Mixed Methods: The Case of Business Environment Promoting Innovation Vitality, Synergizing New Quality Productivity, and "Employment First" [J]. Management World, 2024, 40(06): 217-237. Zhang Liqiong. High-Quality Development of Public Services, Disruptive Innovation, and Common Prosperity in Urban and Rural Areas [J]. Statistics and Decision Making, 2025, 41(11): 95-99. Hu Hongbin. The Driving System, Operational Mechanism, and Improvement Pathways for Advancing Education Powerhouse Construction in the New Era: A Systematic Analytical Framework [J]. Journal of Southwest University for Nationalities (Humanities and Social Sciences Edition), 2024, 45(07): 186-195. Sun Daming, Hu Sumin, Zhu Tianyi, et al. How Industrial Internet Platforms Drive New Quality Productivity Development: A Data Element Perspective [J]. Science and Technology Progress and Policy, 2025, 42(03): 38-49. Shan Wei, Xie Wenbin, Sun Yizhong, et al. Breakthrough Innovation and High-Quality Development of Low-Altitude Economy Enterprises[J/OL]. Studies in Science of Science, 1-26[2025-11-04]. Hu Hongbin. The Driving System, Operational Mechanism, and Improvement Pathways for Advancing Education Powerhouse Construction in the New Era: A Systematic Analytical Framework [J]. Journal of Southwest University for Nationalities (Humanities and Social Sciences Edition), 2024, 45(07): 186-195. Garcia-Castro, Roberto, and Miguel A. Ariño. "A general approach to panel data set-theoretic research." Journal of Advances in Management Sciences & Information Systems 2.63-76 (2016): 526. Zhang, Jichang, Long, Jing, and Wang, Zemin. "What Kind of Institutional Environment Promotes High Entrepreneurial Activity? A Dynamic QCA Analysis Based on Provincial Panel Data." Science and Technology Progress and Policy, 2024, 41(24): 36-48. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8350814","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":565458501,"identity":"1013a1c4-f6ed-4822-83b6-edfe81b1277f","order_by":0,"name":"Chunlin Xiong","email":"","orcid":"","institution":"Hunan Agricultural University College of Public Administration and Law","correspondingAuthor":false,"prefix":"","firstName":"Chunlin","middleName":"","lastName":"Xiong","suffix":""},{"id":565458506,"identity":"1749f845-30eb-4bdc-ad6a-e226e744de2c","order_by":1,"name":"Ren Fan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA50lEQVRIiWNgGAWjYBACNvmD7T8/GPzj4WdvPvggoaKGsBY+CeYGaYmKA3KSPceSDR6cOUZYi5wEe4MEz5kDxgYzcswkH7YwE+Ew6cYGA8m2O4kbJHLMKhIb2Bj427sT8GuROdiQUNj2LHE7z7OyG4k7ZBgkzpzdgF8LQ2LDAck25sSd7cnbbiSeYWMwkMglqKWxgReoZcOBBLOCxDZmIrRIJDYz8Jw5bGxwIsWMgTgtPAfbmCUq0sCBLJFw5hgPQb/It7c/Y/xgYAOOyo8/Kmrk+Nt78WvBADykKR8Fo2AUjIJRgBUAAI5KTv8yzfxSAAAAAElFTkSuQmCC","orcid":"","institution":"Hunan Agricultural University College of Public Administration and Law","correspondingAuthor":true,"prefix":"","firstName":"Ren","middleName":"","lastName":"Fan","suffix":""},{"id":565458507,"identity":"617d4bf2-40ce-4015-8345-cf12afbeb12d","order_by":2,"name":"Yi Liu","email":"","orcid":"","institution":"Hunan Agricultural University College of Public Administration and Law","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Liu","suffix":""}],"badges":[],"createdAt":"2025-12-13 07:08:45","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8350814/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8350814/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99313410,"identity":"324930ac-511d-4844-9798-a0876206367b","added_by":"auto","created_at":"2025-12-31 16:20:07","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":561052,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/19186787cc9dcced4a6d4597.docx"},{"id":99313847,"identity":"07384c75-151a-4bc4-ad67-ab78a7c406d3","added_by":"auto","created_at":"2025-12-31 16:20:32","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":389744,"visible":true,"origin":"","legend":"","description":"","filename":"Figure.docx","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/7f00b66f0c820d48a02d2e39.docx"},{"id":99313356,"identity":"45a9ace8-46a1-44f3-90c0-23a1bbfc4121","added_by":"auto","created_at":"2025-12-31 16:20:03","extension":"json","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5605,"visible":true,"origin":"","legend":"","description":"","filename":"e71f3e586cc94044bfb215d8740f7f8c.json","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/49c59c495119bb791cf5845f.json"},{"id":99313485,"identity":"ff548b0b-9b21-44cf-a252-eaab6492cd7d","added_by":"auto","created_at":"2025-12-31 16:20:14","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":132626,"visible":true,"origin":"","legend":"","description":"","filename":"e71f3e586cc94044bfb215d8740f7f8c1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/652703cfceb3851efe03bf9e.xml"},{"id":99037662,"identity":"bf8e3f98-1370-4497-a465-602a21a85fdd","added_by":"auto","created_at":"2025-12-26 10:44:44","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":264240,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/ca2b172a47e724a5da56acf5.png"},{"id":99314492,"identity":"01495092-2813-45d7-9cb5-db2b68c889a9","added_by":"auto","created_at":"2025-12-31 16:21:35","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102092,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/62494280bc994e9855efb654.png"},{"id":99313522,"identity":"b57825e5-adeb-4450-b1c6-485a6e8e8b06","added_by":"auto","created_at":"2025-12-31 16:20:15","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":148509,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/25e338e1cdb872f983dd58e7.png"},{"id":99314068,"identity":"25e0311d-2253-48f4-a468-0ae573379995","added_by":"auto","created_at":"2025-12-31 16:20:48","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":467582,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/de16bafcca97f77a2ccb0706.jpeg"},{"id":99313084,"identity":"9d233219-a14e-4a38-9734-720f23238ad3","added_by":"auto","created_at":"2025-12-31 16:19:46","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":241163,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/07fc0f2196b55ec4a72c4e14.jpeg"},{"id":99314591,"identity":"a6c72694-2e9a-46c3-9a2f-1eb2ca1a3e1c","added_by":"auto","created_at":"2025-12-31 16:21:55","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":872496,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/fec5067d57c3623b2af2497b.jpeg"},{"id":99037656,"identity":"f231b76b-fbb8-4e8f-ac08-ebee3836c246","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":65948,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/76e384c56eaaabb40828b6f6.png"},{"id":99314455,"identity":"236fbb5b-eae2-4191-8883-14fbe26353c2","added_by":"auto","created_at":"2025-12-31 16:21:31","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21706,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/7de53837b12293436c3fd8a1.png"},{"id":99037664,"identity":"377aecdc-4849-4f7d-b7e7-4f133e4b6d20","added_by":"auto","created_at":"2025-12-26 10:44:45","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":50699,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/190f3c0ef9de2ab7267e47b8.png"},{"id":99314535,"identity":"0780631e-8094-4666-994f-233497162c49","added_by":"auto","created_at":"2025-12-31 16:21:46","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":89284,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/db470d7448aee0647a868948.png"},{"id":99037659,"identity":"4f3961e9-38e1-40ad-9c95-f5390bdd8869","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41895,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/1c775d9b7c2303de314b0ca2.png"},{"id":99037657,"identity":"2fdcbadb-e776-48e3-8102-5909b28e3d4d","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":199315,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/5cd855a960c6be6abe83998f.png"},{"id":99037660,"identity":"51b4ce5c-8f96-4491-b71f-a6794a165c1e","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":131431,"visible":true,"origin":"","legend":"","description":"","filename":"e71f3e586cc94044bfb215d8740f7f8c1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/460f03d369c26974736b70b1.xml"},{"id":99037661,"identity":"373f7715-20e7-4b38-ab33-c50fcb8d4906","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":143262,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/b27f833a1f3310804cf73079.html"},{"id":99037665,"identity":"65d6d022-1916-427b-9a74-5ee8a9a44013","added_by":"auto","created_at":"2025-12-26 10:44:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":111275,"visible":true,"origin":"","legend":"\u003cp\u003eTheoretical Framework Diagram\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/4fb2621c27ccb41d2e282328.png"},{"id":99314018,"identity":"d9738e42-6990-47e7-b1e8-d2b6251a0c54","added_by":"auto","created_at":"2025-12-31 16:20:46","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":71853,"visible":true,"origin":"","legend":"\u003cp\u003eComplex Mediation Model\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/336c77278f50f5a5f716082c.png"},{"id":99037646,"identity":"89ac669f-8e5c-47f2-9eff-0964593c4305","added_by":"auto","created_at":"2025-12-26 10:44:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":204475,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of Mixed-Methods Approach for Testing Complex Mediating Effects\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/b4d4adacf2061e538a9260a0.png"},{"id":109761231,"identity":"6d4e8db2-ebcd-4d58-8f2e-9df79f9933eb","added_by":"auto","created_at":"2026-05-22 07:29:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":747690,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8350814/v1/16e62250-0b0e-4aa4-bb8b-c125b19b6fd9.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"How Does Artificial Intelligence Drive the Optimization of Public Services Structural? ——Complex Intermediary Mechanisms Based on the Digital Ecosystem","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIn recent years, the rapid global advancement of artificial intelligence (AI) technology has catalyzed profound transformations in public service delivery. Recognizing AI's strategic importance, numerous countries have accelerated the formulation and implementation of dedicated policies. In 2024, the European Union adopted the Artificial Intelligence Act, establishing foundational regulatory safeguards for AI applications in high-risk domains such as public administration. According to the Organization for Economic Cooperation and Development (OECD), over 70% of its member states have integrated AI into their digital government strategies, yielding significant improvements in critical public service sectors including healthcare, education, and transportation. In China, the \"AI+Public Services\" initiative was formally proposed in the 2025 Guidelines on Deepening the Implementation of the \"AI+\" Initiative. As the \"Digital China\" strategy continues to advance, AI—positioned as a pivotal driver of the new wave of technological revolution and industrial transformation—plays an increasingly central role in addressing systemic challenges in public services, including insufficient coordination\u003csup\u003e1\u003c/sup\u003e, technical adaptation barriers\u003csup\u003e2\u003c/sup\u003e, information asymmetry\u003csup\u003e3\u003c/sup\u003e, and unequal resource distribution\u003csup\u003e4,5\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eHowever, empirical evidence indicates that isolated, single-purpose applications of AI technology can achieve only localized efficiency gains and are largely insufficient to drive systemic and holistic quality improvements in public service systems\u003csup\u003e6\u003c/sup\u003e. These limitations are manifested across three primary dimensions: First, the pervasive existence of \"data silos\" impedes cross-departmental and cross-level data sharing, thereby undermining the integrated effectiveness of AI deployment. Second, a persistent misalignment between technological research and development (R\u0026amp;D) and actual public service demands results in solutions that fail to align precisely with real-world contexts, weakening their capacity to generate tangible value. Third, a critical shortage of interdisciplinary talent—professionals equipped with expertise in both AI and public administration—has emerged as a key bottleneck constraining the deep integration and sustainable application of digital technologies in governance.\u003c/p\u003e\n\u003cp\u003eThus, the essential catalyst for transforming public services is the construction of a synergistic and high-performance digital ecosystem, premised on the deep embedding of technology within practical operational contexts. This foundation supports the progressive maturation of applications across critical public service domains, including government approvals, smart healthcare, and accessible education. Consequently, the pivotal issues demanding immediate scholarly and policy attention revolve around elucidating the pathways through which artificial intelligence drives the optimization of public service frameworks, and defining the architecture of a digital ecosystem capable of fully activating AI's potential, overcoming the \"digital divide,\" and thereby realizing structural optimization in public services.\u003c/p\u003e\n\u003cp\u003eIn light of this, the present study combines configuration theory with digital ecosystem theory, employing a hybrid approach of dynamic QCA and regression analysis. A sample of 30 provinces in China (excluding Tibet) from 2019 to 2023 was utilized to overcome the limitations of traditional research frameworks that exclusively prioritize technology application. Concurrently, a systematic analytical framework encompassing five dimensions—innovation, green development, openness, sharing, and coordination—is being constructed to explore the complex causal mechanisms through which AI influences public service structural optimization within the combination of digital ecosystem elements. This provides empirical evidence or theoretical frameworks and practical insights for the digital and intelligent development of public services.\u003c/p\u003e\n\u003cp\u003eThis study makes significant marginal contributions theoretically and methodologically. Theoretically, it introduces an innovative perspective that synthesizes AI's direct influences with a nuanced analysis of the digital ecosystem's mediating functions, thereby challenging the prevalent technological determinism in public administration scholarship. Methodologically, it pioneers a novel approach by synthesizing dynamic QCA with complex mediation analysis. This synthesis adeptly elucidates the multiple concurrent causal configurations whereby AI impacts public services through the digital ecosystem, furnishing a novel analytical framework for interrogating complex causal mechanisms in social sciences.\u0026nbsp;\u003c/p\u003e"},{"header":"Literature Review","content":"\u003cp\u003e\u003cstrong\u003e1.1 Research on Artificial Intelligence and Public Services\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn recent years, the application of artificial intelligence (AI) to transform and enhance public services has become a prominent focus in academic research. This body of scholarship has primarily centered on three interrelated domains: AI applications in public service delivery, the transformation of government governance models through AI, and AI-enabled mechanisms for public feedback.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFirst, artificial intelligence and public service applications. Current scholarly research primarily treats AI as a technological enabler applied across specific service domains, organized around three key dimensions. First, intelligent interaction and automation\u003csup\u003e7\u003c/sup\u003e: generative AI and chatbots are widely deployed in government hotlines, tax consultation, legal aid, and related services, enabling 24/7 digital responsiveness and significantly improving service accessibility and response speed\u003csup\u003e8\u003c/sup\u003e. Second, data-driven decision-making: big data analytics and machine learning algorithms are leveraged to anticipate public demand and optimize resource allocation\u003csup\u003e9,10\u003c/sup\u003e. For example, in smart healthcare, AI supports clinical diagnosis and epidemic forecasting; in smart transportation, it facilitates real-time traffic management and accident prediction\u003csup\u003e11\u003c/sup\u003e. Third, technology-enabled empowerment and security: emerging technologies such as blockchain enhance the transparency, traceability, and integrity of government data\u0026nbsp;\u003csup\u003e4\u003c/sup\u003e. In critical areas including welfare distribution, government procurement, and identity verification, these tools help strengthen public trust\u003csup\u003e12\u003c/sup\u003e. Collectively, these AI applications not only streamline service delivery processes but also reduce administrative burdens\u003csup\u003e13\u003c/sup\u003e, thereby enhancing both operational efficiency and service quality.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSecond, artificial intelligence and the transformation of government governance models. The application of artificial intelligence has profoundly reshaped the organizational logic and operational patterns of governments, accelerating the modernization of governance systems. First, it drives a shift in organizational structures from traditional bureaucratic hierarchies toward flat, networked, and data-driven configurations. This transition necessitates dismantling entrenched data silos to enable efficient cross-departmental collaboration\u003csup\u003e14\u003c/sup\u003e. Second, it enhances scientific decision-making\u003csup\u003e15\u003c/sup\u003e. AI provides robust capabilities in simulation, scenario analysis, and impact assessment for policy formulation, facilitating a transition from experience-based judgment to data-driven, precision-oriented analysis\u003csup\u003e16\u003c/sup\u003e. Third, it fosters the emergence of a new ecosystem of multi-stakeholder collaborative governance, in which governments, enterprises, social organizations, and citizens form synergistic partnerships through AI-powered platforms\u003csup\u003e17\u003c/sup\u003e. However, this transformation also presents significant ethical challenges\u0026mdash;including algorithmic fairness, accountability mechanisms, and data privacy\u0026mdash;that demand the urgent development of comprehensive governance frameworks\u003csup\u003e18\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThird, AI and public feedback. The ultimate effectiveness of AI applications is reflected in citizens\u0026rsquo; sense of fulfillment, satisfaction, and the generation of social value. On one hand, AI-driven personalized services can precisely align with citizens\u0026rsquo; diverse needs\u003csup\u003e19\u003c/sup\u003e, thereby significantly improving public service experiences and user satisfaction\u003csup\u003e20\u003c/sup\u003e. On the other hand, careful attention must be paid to the risks posed by the \u0026quot;digital divide\u0026quot;\u003csup\u003e21\u003c/sup\u003e. In the absence of inclusive design, AI applications may marginalize vulnerable groups such as the elderly and persons with disabilities, potentially exacerbating existing social inequalities. Therefore, advancing inclusive design requires ensuring that technological progress achieves genuine universality, enabling equitable access and participation\u0026mdash;conditions essential for translating innovation into enhanced social welfare and strengthened government credibility\u003csup\u003e22\u003c/sup\u003e, which constitutes the core of public value creation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;1.2 The Systematic Evolution of Digital Technology into a Digital Ecosystem\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn recent years, academic inquiry has predominantly focused on the application pathways of artificial intelligence in public services\u003csup\u003e23,24\u003c/sup\u003e. However, much of this scholarship has treated AI as a homogeneous technological instrument, emphasizing its general utility and outcomes\u0026mdash;resulting in fragmented rather than systematic solutions. A deeper analysis reveals a significant evolution: digital technologies are increasingly coalescing into an integrated digital ecosystem. This transformation redefines technology\u0026rsquo;s role from a mere tool for enhancing efficiency to a structural force capable of fundamentally reshaping patterns of production and daily life, models of governance, and even broader civilizational frameworks. Consequently, this shift extends beyond the deployment of discrete tools toward the orchestrated interaction of multiple interconnected components within a complex, adaptive system.\u003c/p\u003e\n\u003cp\u003eClassical systems theory outlines an \u0026quot;input-integration-output-feedback\u0026quot; paradigm as a foundational framework for understanding complex system operations\u003csup\u003e25,26\u003c/sup\u003e. This paradigm describes a process whereby external resources are transformed into effective outputs through internal structural processing. Its high level of abstraction provides a macro-level, process-oriented lens for analyzing the intricate ecosystems of the digital age. Digital ecosystems themselves have conceptual origins in the interdisciplinary adaptation and reimagining of the traditional \u0026quot;ecosystem\u0026quot; concept. Their theoretical roots can be traced to Moore\u0026apos;s (1993) \u0026quot;business ecosystem\u0026quot; theory, which emphasizes the co-evolutionary capacity of interconnected organizations within an economic community. With the advent of the digital era, a distinct theoretical perspective known as the \u0026quot;digital ecosystem\u0026quot; has emerged. In China, this concept was formally elevated to a policy level in 2021 when the State Council\u0026apos;s \u0026quot;14th Five-Year Plan and the Long-Range Objectives Through 2035\u0026quot; called for establishing a digital rules system to foster an open, healthy, and secure digital ecology. Qiu et al. (2025) further proposed that the core dimensions of digital ecosystem theory encompass four interrelated aspects: interconnected elements, actors, interdependent interactions, and developmental evolution. They characterize digital ecosystems as complex systems that maintain relative independence while interacting with natural and social ecosystems.\u003c/p\u003e\n\u003cp\u003eIn light of this, this paper develops a systematic \u0026quot;technology input\u0026quot; framework, which is realized through the introduction of emerging digital technologies such as artificial intelligence, serving as the initial driver for public service transformation\u003csup\u003e27\u003c/sup\u003e. However, the realization of technological efficacy depends critically on the \u0026quot;system integration\u0026quot; phase. This phase entails complex coordination and configurational interactions across the five dimensions of the digital ecosystem: digital infrastructure, digital government, digital economy, digital society, and digital capabilities. Technological elements achieve value transformation through deep coupling with the diverse actors, data resources, institutional rules, and application scenarios embedded within these dimensions. This process fundamentally reshapes the underlying logic of public service delivery\u003csup\u003e28\u003c/sup\u003e, transforming it from a traditional supply-oriented model into value-added outcomes that are co-created by multiple stakeholders\u0026mdash;including governments, enterprises, and the public\u0026mdash;within a digital environment\u003csup\u003e29\u003c/sup\u003e. At the core of this model is artificial intelligence as a technological input. Its potential must be effectively activated and regulated through differentiated configurations of the digital ecosystem as the system integrator, ultimately converting technological potential into tangible \u0026quot;value outputs.\u0026quot; These outputs are manifested as structural optimizations of public services across five dimensions: innovation, green development, openness, sharing, and coordination. The theoretical framework is illustrated in Figure 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2 Theoretical Mechanisms and Research Pathways\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.1 Digital Ecosystem and Public Service Structural Optimization: A Configuration Perspective\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the current era of global digital transformation, building a robust digital ecosystem has become the key driver behind the iterative upgrading of public service structures. However, the digital ecosystem is not the result of a single isolated factor, but rather a complex system comprising five interrelated dimensions: digital infrastructure, digital economy, digital government, digital society, and digital capability. These dimensions are deeply intertwined and mutually reinforcing. Through diverse configurations, they act individually or collectively across different value dimensions of public service structural optimization, jointly shaping the mode, quality, and efficiency of public service delivery and ultimately enabling the co-creation of public value.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(1) Digital infrastructure, as the foundational backbone supporting public service development, comprises robust network connectivity, digital hubs, and computing power\u0026mdash;collectively forming the underlying architecture for efficient public service operations\u003csup\u003e30\u003c/sup\u003e. High-speed networks facilitate the seamless flow of data and enable cross-domain collaboration, while the intensive use of cloud computing significantly enhances resource utilization efficiency and supports the green and low-carbon transformation of public services. Notably, the rapid advancement of computing systems provides core technical support for complex public service applications such as smart cities. Therefore, building advanced, inclusive, and secure digital infrastructure is an indispensable prerequisite for enhancing the quality of public service delivery.\u003c/p\u003e\n\u003cp\u003e(2) As the most dynamic component, the digital economy serves as a powerful external driver and source of innovation for optimizing public service structures. Emerging technologies and business models\u0026mdash;including big data, artificial intelligence, and platform economies\u0026mdash;are increasingly integrated into public services, giving rise to innovative service paradigms such as smart healthcare and personalized education\u003csup\u003e31\u003c/sup\u003e. The principles of the sharing economy not only optimize social resource allocation and improve overall operational efficiency but also reflect the core values of green development. Leading technology enterprises, through public-private partnerships, open their mature core capabilities\u0026mdash;such as payment and identity authentication systems\u0026mdash;to government agencies, significantly improving service convenience and diversity. Moreover, the vigorous growth of the digital economy compels governments to accelerate data openness, unlocking the social value of public data and fostering a virtuous cycle of \u0026quot;building through utilization\u0026quot; in public service optimization.\u003c/p\u003e\n\u003cp\u003e(3) Digital government construction functions as the core engine for public service structural optimization, fundamentally reshaping service delivery models through deep institutional transformation. Institutionally, digital government promotes frameworks such as \u0026quot;one-stop online services\u0026quot; and \u0026quot;integrated online governance,\u0026quot; effectively addressing service fragmentation by enabling cross-departmental and cross-level coordination of business processes and data sharing. Simultaneously, digital government actively releases public data to society via established open platforms\u003csup\u003e32\u003c/sup\u003e, safeguarding citizens\u0026rsquo; rights to information and oversight while stimulating social innovation. The comprehensive digitization of government services overcomes traditional geographical constraints, significantly advancing the equalization and universal access to basic public services. Overall, the depth and breadth of digital government development directly determine the extent to which public service efficiency can be enhanced.\u003c/p\u003e\n\u003cp\u003e(4) The maturity of a digital society directly reflects citizens\u0026rsquo; digital living standards and participatory capacities, serving as a critical factor in determining whether public service reforms achieve broad societal acceptance and effective implementation. In a highly developed digital society, individuals skillfully employ digital tools for expression, oversight, and interaction. This widespread civic engagement shifts public service provision from a \u0026quot;government-led\u0026quot; model toward a \u0026quot;people-centered\u0026quot; approach, enabling more responsive alignment with genuine public needs. Public wisdom and creativity emerge as vital sources of innovation, exemplified by crowdsourced models in environmental monitoring and community governance. A robust digital social ecosystem further strengthens mutual trust between government and citizens, contributing to broader societal cohesion.\u003c/p\u003e\n\u003cp\u003e(5) Digital capability centers on the human element, functioning as the soft power and fundamental enabler for realizing public service structural optimization. It encompasses both citizens\u0026rsquo; digital literacy and civil servants\u0026rsquo; digital competencies\u003csup\u003e33\u003c/sup\u003e. On one hand, improving nationwide digital literacy is essential for bridging the digital divide and ensuring equitable access to the benefits of digital development across all population groups; otherwise, technological progress may deepen existing inequalities. On the other hand, civil servants equipped with digital thinking and execution skills serve as internal catalysts, empowering governments to proactively adopt new technologies for process reengineering and service innovation. Strong digital capabilities also facilitate effective communication and collaboration among diverse governance actors, promoting coordinated and integrated service delivery. Thus, digital capacity building constitutes the decisive factor in activating the entire digital ecosystem, ensuring that public service reforms remain people-centered and sustainable.\u003c/p\u003e\n\u003cp\u003eIn summary, the five dimensions of the digital ecosystem do not operate in isolation. Rather, they interact through interconnected and combinatory mechanisms, synergistically driving the optimization of public service structures. This paper examines the integrated pathways through which the digital ecosystem enhances public service structural reform from a configurational perspective, laying the theoretical foundation for further exploration of the relationships and influence mechanisms among artificial intelligence, the digital ecosystem, and public service structural optimization.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Complex Mediation Effect: From Configuration Pathways to Mechanism Testing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2.1 Direct Effects\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAs a critical technological input, artificial intelligence provides essential enabling capabilities for optimizing public service structures through its advanced computing power and sophisticated algorithms. First, at the innovation level, AI drives innovation in service models and operational formats, enabling predictive regulation by analyzing large-scale risk data to issue early warnings and implement timely interventions prior to safety incidents. It also supports personalized public services, reflecting a transformative shift from \u0026quot;standardization\u0026quot; to \u0026quot;diversification\u0026quot;\u0026mdash;a set of revolutionary advancements enabled by AI\u003csup\u003e34\u003c/sup\u003e. Second, in terms of sustainability, AI\u0026rsquo;s contributions extend well beyond resource optimization. Through millisecond-level monitoring, prediction, and simulation of urban energy systems, building energy consumption, and traffic flows, AI enables precise supply-demand matching and dynamic scheduling while significantly reducing wasteful energy use. For instance, AI-powered smart grids balance fluctuations in renewable energy generation, while intelligent transportation systems effectively minimize emissions caused by traffic congestion, making the operation of public service systems themselves exemplars of green, low-carbon efficiency. Third, at the openness level, AI enhances government transparency and facilitates public participation. By leveraging data mining and visualization tools, AI transforms raw, complex datasets into intuitive charts, actionable insights, and interactive applications. This substantially lowers the barriers for citizens to understand and utilize open government data, thereby enhancing its practical utility and fostering greater civic engagement. Fourth, at the sharing level, AI can intelligently analyze massive, multi-source datasets to more accurately forecast trends, identify risks, and optimize resource allocation. This enables public service providers to precisely identify the needs of diverse regions and population groups, directing resources toward grassroots communities, rural areas, and vulnerable populations while minimizing misallocation and waste. For example, AI can predict changing demands for public transportation, elderly care facilities, and logistics networks, guiding public investment with greater spatial precision. This improves the equitable distribution of urban-rural infrastructure and services, promoting coordinated development from the planning stage onward. In summary, this paper proposes Hypothesis 1:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHypothesis H1:\u003c/strong\u003e Artificial intelligence has a significant positive impact on the optimization of public service structures.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2.2 Indirect Effects\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe key to AI serving the public sphere lies in the rational construction of digital ecosystems. As a vital technological input, AI\u0026rsquo;s functionality depends on the systematic integration of digital services and technological ecosystems to achieve the ultimate goal of delivering high-quality public services. This implies that digital ecosystems serve as an essential mediating mechanism between AI and public service structural optimization, manifested specifically in three critical challenges: First, algorithmic dependency and inclusivity challenges. The effective operation of AI relies on high-quality, large-scale data. If digital infrastructure coverage is uneven or public digital literacy is inadequate\u003csup\u003e35,36\u003c/sup\u003e, AI may exacerbate the \u0026quot;digital divide,\u0026quot; thereby undermining equitable access to public services. Second, algorithmic bias and fairness challenges. AI systems can replicate and amplify existing societal biases during automated decision-making, raising serious ethical concerns that threaten equity in public service delivery\u003csup\u003e37\u003c/sup\u003e. Third, black-box decision-making and accountability challenges. The opacity of AI-driven decision processes poses significant difficulties for traditional government accountability mechanisms, creating a tension between data-intensive governance and bureaucratic accountability structures. In summary, without robust ecosystem support, AI applications risk generating new crises in fairness, ethics, and governance\u0026mdash;even as they enhance operational efficiency.\u003c/p\u003e\n\u003cp\u003eTherefore, adopting a systemic perspective through the lens of the digital ecosystem is essential to mitigate risks associated with AI applications and ensure the realization of public value. As an organic whole, the digital ecosystem provides indispensable support for the healthy and sustainable development of AI. Robust digital infrastructure ensures widespread access to AI services and enables energy-efficient computing; strong digital capabilities constitute fundamental prerequisites for bridging the digital divide and facilitating broad AI adoption; effective digital governance establishes institutional mechanisms such as algorithmic auditing and ethical oversight frameworks to systematically address bias and safeguard fairness. A vibrant digital economy and a mature digital society collectively provide AI with a continuous source of innovation and broad-based societal acceptance. Thus, the digital ecosystem is not passive but functions as a dynamic intermediary\u0026mdash;capable of being activated by AI while simultaneously regulating and guiding its deployment. In summary, this paper proposes Hypothesis 2: \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHypothesis H2:\u0026nbsp;\u003c/strong\u003eThe digital ecosystem plays a pivotal intermediary role in AI-driven optimization of public service structures, where varying configurations of its components lead to distinct pathways and outcomes.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2.3 Research Framework and Core Questions\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eGiven this, this paper develops a complex intermediary model from the perspective of the digital ecosystem, examining how AI technology interacts with the digital ecosystem to influence public service structural optimization. It seeks to address the following research questions: How does AI directly impact public service structural optimization? What configurations of the digital ecosystem can effectively drive this transformation? Furthermore, as a complex intermediary, how do different combinations of its internal elements differentially shape the translation of technological inputs into high-quality public service outputs? The complex intermediary model proposed in this paper, illustrated in Figure 2, clearly delineates the complete causal pathway from technological inputs to value outcomes.\u003c/p\u003e"},{"header":"Research Design","content":"\u003cp\u003e\u003cstrong\u003e3.1 Complex Mediating Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConfronted with multiple concurrent causal relationships in research on public service structural optimization, traditional single statistical methods struggle to uncover deep-seated interdependencies among variables. Therefore, drawing on the approach of Du et al. (2024), this study integrates dynamic qualitative comparative analysis (QCA) with regression analysis to construct a mixed-methods testing framework\u003csup\u003e38\u003c/sup\u003e. This framework aims to identify multiple equivalent configuration pathways through which the digital ecosystem influences public service structural optimization, while simultaneously assessing the indirect effects of artificial intelligence on public service structural optimization via these mediating configurations.\u003c/p\u003e\n\u003cp\u003eThe complex mediation model comprises four steps (see Figure 3): First, examine the M-Y causal chain. Employ dynamic QCA analysis to identify multiple condition combinations influencing the outcome variable\u0026mdash;specifically, how different configurations of the digital ecosystem affect the optimization of public service structures. Second, assign values to configuration M, converting dynamic QCA results into mediating variables for regression analysis. Third, test the X-M causal chain to examine the overall effect of AI on the formation of the digital ecosystem (entropy values across five dimensions), i.e., the effect of path a. Step 4: Test the X-Y causal chain to further explore AI\u0026apos;s indirect influence through digital ecosystem configurations. Key notes: ① During configuration membership calculation, a case\u0026apos;s membership in configuration M is determined by taking the minimum membership value across all condition sets; ② When multiple configuration elements are interdependent, separate regression analyses treat different configurations as independent mediating variables to avoid multicollinearity.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Data Sources\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study utilizes a panel of 30 Chinese provinces (excluding Tibet) covering the period from2019 to 2023 to examine the complex interplay among artificial intelligence, the digital ecosystem, and the structural optimization of public services. Data on AI and the digital ecosystem are drawn from the Digital Ecology Index. Additional variables are sourced from a comprehensive set of official publications, including the China Statistical Yearbook, Health Statistical Yearbook, China Local Digital Service Capability, China Fiscal Yearbook, China Government Transparency Report, China Environmental Statistical Yearbook, China Industrial Statistical Yearbook, and the China City Statistical Yearbook, supplemented by provincial statistical yearbooks and data from the National Bureau of Statistics. Missing values in the dataset were handled using linear interpolation and mean imputation methods to\u0026nbsp;ensure completeness.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3 Data and Measurement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.3.1 Dependent Variable: Artificial Intelligence\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eGiven that artificial intelligence is a multidimensional, pervasive general-purpose technology, it is challenging to comprehensively measure its overall development level and application potential at the provincial level using a single absolute indicator. Therefore, this study adopts the \u0026quot;Digital Innovation\u0026quot; sub-index from the 2019-2023 Digital Ecosystem Index, jointly released by the National Engineering Laboratory for Big Data Analysis and Application Technology at Peking University and multiple institutions, as a metric for assessing AI development.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.3.2 Explanatory Variable:\u0026nbsp;\u003c/em\u003e\u003cem\u003eOptimization of public service structure\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePublic service structure optimization is defined as advancing the public service system from \u0026quot;ensuring basic coverage and broad accessibility\u0026quot; toward \u0026quot;enhancing quality, expanding capacity, and achieving balanced accessibility,\u0026quot; with the fundamental goal of meeting the people\u0026apos;s growing needs for a better life. This ultimately aims to realize a new stage of development characterized by higher levels, higher quality, greater efficiency, greater fairness, and greater sustainability. Therefore, drawing on Zhang (2025)\u003csup\u003e39\u003c/sup\u003e, this paper constructs an indicator system for optimizing public service structures based on development levels across five dimensions: innovation, green development, openness, sharing, and coordination (see Table 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 1: Indicator System for Public Service Structure Optimization\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"562\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePrimary Indicators\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eSecond-Level Indicators\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eMeasurement Method\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003eAttribute\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePublic Service Innovation Development Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eHealthcare Talent Service Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eNumber of licensed physicians, registered nurses, and pharmacists / Total population\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eDigital Education Service Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eNumber of terminals in public library e-reading rooms\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePublic Service Green Development Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eGreen coverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eGreen coverage rate in built-up areas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eEnvironmental Protection Investment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eEnvironmental Protection Expenditures / General Public Budget Expenditures\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eWaste treatment level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003ePublic Household Waste Harmless Treatment Rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePublic Service Open Development Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eGovernment Service Openness Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eGovernment Transparency Index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eLevel of Trade in Services Openness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eLevel of Trade in Services Openness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePublic Service Shared Development Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eLevel of Shared Healthcare Services\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eNumber of Beds in Public Healthcare Institutions / Total Population\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eLevel of Shared Educational Services\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003ePublic library collection size per total population\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e+\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003ePublic Service Coordinated Development Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eLevel of Coordination in Urban and Rural Elderly Care Services\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003eNumber of Urban Pension Insurance Participants / Number of Rural Pension Insurance Participants\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eLevel of Coordination in Urban and Rural Transportation and Communications\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003ePer capita transportation and communication expenditure of urban residents / Rural residents\u0026apos; transportation and communication expenditure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003eUrban-Rural Per Capita Healthcare Coordination Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 229px;\"\u003e\n \u003cp\u003ePer capita healthcare expenditure of urban residents / Per capita healthcare expenditure of rural residents\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 71px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cem\u003e3.3.3 Mediating Variable: Digital Ecosystem\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSince the Digital Ecosystem Index report aggregates underlying indicators into standardized indices through its proprietary weighting system, this study directly adopts the published final sub-index values. Based on the Digital Ecosystem Index framework, the digital ecosystem is categorized into five dimensions:\u0026nbsp;①\u0026nbsp;Digital Infrastructure, encompassing infrastructure development, data resources, and policy environment;\u0026nbsp;②\u0026nbsp;Digital Government, including the Online Government Service Capability Index, Smart Environmental Protection Index, Digital Government Development Index, and Rural Digital Governance Index;\u0026nbsp;③\u0026nbsp;Digital Economy, comprising the Big Data Industry Development Index, Artificial Intelligence Industry Development Index, Digital Industry Electricity Consumption Index, Digital Economy Investor Confidence Index, Enterprise Digital Transformation Index, SME Digitalization Index, Micro and Small Enterprise Digital Development Index, and Rural Digital Economy Index;\u0026nbsp;④\u0026nbsp;Digital Society, covering the Digital Inclusive Finance Index, Digital Lifestyle Index, Social Dispute Search Index, Digital Convenience Payment Index, and Rural Digital Society Index;\u0026nbsp;⑤\u0026nbsp;Digital Capability, consisting of Digital Talent and Digital Security. These five dimensions collectively constitute an integrated system that comprehensively reflects the overall state of a region\u0026rsquo;s digital ecosystem.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.3.4 Control Variables\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo prevent omitted variable bias from affecting estimation results, a set of control variables was incorporated based on established theoretical and empirical research\u003csup\u003e40,41\u003c/sup\u003e. ① Government intervention, measured by the ratio of local public budget expenditure to GDP; ② Urbanization rate, calculated as the proportion of urban population to total population; ③ Industrial structure, captured by the share of tertiary industry value-added in GDP; ④ Level of financial development, measured by the sum of deposit and loan balances of financial institutions as a percentage of GDP.\u0026nbsp;\u003c/p\u003e"},{"header":"Empirical Analysis Results","content":"\u003cp\u003e\u003cstrong\u003e4.1 Configural Analysis of Digital Ecosystems Optimizing Public Service Structure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo ensure consistency in calibration standards over time and enhance the longitudinal comparability of results, this study employs a theory-driven dynamic QCA fixed-anchor calibration approach\u003csup\u003e42\u003c/sup\u003e, setting threshold values for full membership, cross-case intersections, and full non-membership at the 75th percentile, 50th percentile, and 25th percentile, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.1.1 Necessity Condition Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePrior to conducting the configurational analysis, the necessity of each antecedent condition must be verified. When the consistency level is greater than or equal to 0.9, the condition variable can be considered a necessary condition for the outcome variable \u003csup\u003e43\u003c/sup\u003e. However, in panel data QCA analysis, consistency-adjusted distance must be further employed to examine whether necessary conditions exhibit temporal or case effects, ensuring the reliability of aggregated consistency. Calculated via formula, a consistency-adjusted distance below 0.2 indicates higher precision in aggregated consistency and coverage\u003csup\u003e44\u003c/sup\u003e. When the adjusted distance exceeds 0.2, researchers should further investigate its necessity\u003csup\u003e45\u003c/sup\u003e. Table 2 shows that the consistency levels for all conditions are below 0.9, indicating no necessary conditions affecting the optimization of high public service structures. To verify the stability of these results, the consistency distance was adjusted. Among the groups\u0026mdash;non-high digital government, high digital society, non-high digital society, non-high digital economy, and non-high digital capability\u0026mdash;the adjusted distances exceeded 0.2. Annual consistency level checks revealed excessively large adjusted distances between variable groups. When consistency levels ranged between 0.1 and 0.9 and inter-group coverage exceeded 0.5, these conditions were deemed non-essential.\u003c/p\u003e\n\u003cp\u003eTable 2 Necessity Analysis Conditions\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"582\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eCondition variable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eHigh-Level Public Service Structural Optimization\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 260px;\"\u003e\n \u003cp\u003eLow-Level Public Service Structural Optimization\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eAggregate Consistency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003eAggregate Coverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eIntergroup Consistency Adjustment Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003eIntra-Group Consistency-Adjusted Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eAggregate Consistency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003eAggregate Coverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003eInter-Group Consistency-Adjusted Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003eIntra-Group Consistency-Adjusted Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eHigh-digit foundation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.808\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.081\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.460\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.348\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.343\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.258\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.690\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eNon-high digit foundation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.326\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.331\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.200\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.719\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.805\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.151\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.311\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eHigh Digital Government\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.788\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.449\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.340\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.510\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.702\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eNon-high digital government\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.337\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.335\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.391\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.785\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.789\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.223\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.276\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eHigh Digital Society\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.759\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.756\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.246\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.460\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.359\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.362\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.617\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.638\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eNon-high digital society\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.359\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.527\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.661\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.758\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.761\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.322\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.293\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eHigh Digital Economy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.842\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.829\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.171\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.518\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.311\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.432\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.748\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eNon-high digital economy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.299\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.300\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.499\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.730\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.828\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.842\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.206\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.270\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eHigh Numerical Ability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.839\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.832\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.104\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.420\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.309\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.304\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.742\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003eNon-High Numerical Ability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.304\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.303\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.252\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.736\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e0.833\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.840\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.090\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 68px;\"\u003e\n \u003cp\u003e0.334\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003e4.1.2 Configurational Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eConfiguration analysis is the distinctive core function of the QCA method, primarily used to reveal the impact of different condition combinations on the outcome variable. Referencing existing research and considering practical circumstances, this study ultimately selected a consistency threshold of 0.85, a case frequency of 2, and a PRI threshold of 0.7. It prioritized intermediate solutions while supplementing with minimal solutions to identify core and peripheral conditions. Table 3 presents the overall configuration analysis results. The aggregated consistency of the high-public-service-structure optimization solution reached 0.916, exceeding the 0.75 threshold. Moreover, the inter-group consistency adjustment distance for each configuration remained below 0.2, while the highest intra-group consistency adjustment distance was 0.212. This indicates strong explanatory power for the aggregated consistency, confirming that these three configurations constitute sufficient conditions for achieving high public service structure optimization. This outcome comprises three configurations: namely Ecologically Mature Type (H1), Administratively Driven Type (H2), and Socially Empowered Type (H3). Subsequently, each configuration undergoes detailed analysis and comparative evaluation.\u003c/p\u003e\n\u003cp\u003e(1) Mature Ecosystem Model (H1). This path relies on digital infrastructure, digital government, digital society, and digital economy as core prerequisites, with its intrinsic momentum stemming from deep inter-system coordination and virtuous cycles. Typical examples of this path include developed provinces and cities like Beijing, Shanghai, Jiangsu, Zhejiang, and Guangzhou. These regions lead China across multiple dimensions\u0026mdash;hardware construction, institutional guidance, industrial vitality, and social applications\u0026mdash;forming a development pattern characterized by comprehensive leadership and balanced elements. Within this system, while citizens\u0026apos; digital capabilities are not explicitly highlighted as a primary condition, their overall high level of development has long become the foundational bedrock for societal operations. The formation of this pathway stems from three key factors: First, top-tier digital infrastructure provides a robust foundation for digitalization across all sectors; second, forward-looking digital governance frameworks and efficient institutional supply offer rule-based guidance and policy safeguards for the coordinated evolution of the entire system; third, vibrant digital economic innovation and widespread digital societal applications have not only spawned new business models and formats but also compelled service optimization and upgrading. Thus，the ecosystem maturity path represents a more advanced developmental form emphasizing systemic synergy，innovation-driven growth，and value-led guidance，serving as a benchmark和model for regional digitalization. Within such an ecosystem，public services continuously enhance quality和efficiency，ultimately maximizing public value realization.\u003c/p\u003e\n\u003cp\u003e(2) Administration-Driven Model (H2). This path centers on digital infrastructure, digital government, digital economy, and digital capabilities as core prerequisites, representing a government-led, efficiency-first development model. While its case distribution overlaps somewhat with the Ecologically Mature Model (H1), its most representative provinces are municipalities like Tianjin and Chongqing. This indicates that even in developed regions with relatively mature digital ecosystems, strong government guidance remains a crucial driver for upgrading public service quality. Leveraging the centralized and efficient advantages of their administrative systems, these regions achieve rapid progress in key areas like government digitalization through top-down design, resource allocation, and targeted policies\u0026mdash;even when the broader social digital application environment is still developing. The effectiveness of this approach hinges on three pillars: first, unified institutional arrangements provide stable expectations and directional guidance for development; second, efficient execution systems and performance incentives ensure the thorough implementation of policy objectives; third, the government\u0026apos;s control over critical resources and infrastructure lays the foundation for coordinated action across the entire region. Consequently，the administratively driven model significantly shortens innovation diffusion cycles in areas requiring rapid response and concentrated resource allocation，achieving a \u0026quot;point-to-area\u0026quot; development effect.\u003c/p\u003e\n\u003cp\u003e(3) Social Empowerment Pathway (H3). This pathway centers on digital infrastructure, digital society, digital capabilities, and digital government as core prerequisites. Unlike traditional economic development models that prioritize growth first, this approach embodies an inclusive and universally accessible path of social empowerment. Examining specific cases, representative provinces for this path include Hebei and Henan. In these regions, the contribution of digital economy industries remains relatively weak, and government efficiency is not high. However, as key areas actively advancing the optimization of public service structures, the socially empowered development path can achieve this goal by strengthening digital infrastructure construction, promoting the widespread application of digital society, and enhancing the digital capabilities of the entire society. The reasons for this path emerging are as follows: first, robust digital infrastructure provides a solid foundation for the development of public services; second, effective digital government design and policy implementation provide institutional safeguards for the digital transformation of public services and the enhancement of public digital literacy; third, widespread digital societal applications and public demand create a synergistic force for the reapplication of digital technologies，thereby compelling improvements in public service quality. Thus，the socially empowered development path represents a model that places greater emphasis on social participation，demand-driven growth，and inclusive sharing.\u003c/p\u003e\n\u003cp\u003eIn summary, Pathways H1, H2, and H3 are not mutually exclusive but exhibit a progressive, convergent, and diffuse evolutionary trend. On one hand, in terms of driving logic, Pathway H1 represents a mature ecosystem that relies on the comprehensive synergy of the five key elements of the digital ecosystem to form a self-sustaining virtuous cycle within the system, embodying a comprehensive, naturally evolved advanced form; Pathway H2 highlights the government\u0026apos;s leading role, concentrating resources on breakthroughs in infrastructure, industrial cultivation, and capacity building for key groups. It leverages administrative power to compensate for deficiencies in the societal digital ecosystem, manifesting as a top-down, efficiency-first intervention model; Pathway H3 emphasizes leveraging digital foundations as a base to drive inclusive public service coverage under relatively weak digital economic momentum. It harnesses dual forces of government guidance and social participation, presenting an inclusive approach characterized by foundational prioritization and demand-driven growth. Trend-wise，Pathway H1 signifies progress from high-level individual elements toward highly efficient synergy. Pathway H2 signifies a gradual shift in the government\u0026apos;s role from builder to regulator and enabler，focusing on rule-setting and platform-building. Pathway H3 enhances the socialization of public services by stimulating digital vitality at the societal level，thereby driving the overall upgrade of digital government and digital industries.\u003c/p\u003e\n\u003cp\u003eTable 3 Configuration Analysis Results\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"548\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eCondition Variable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 387px;\"\u003e\n \u003cp\u003eConfiguration Analysis - High\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003eEcosystem Maturity Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003eAdministrative-Driven Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003eSocially Empowered Type\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDigital Infrastructure\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDigital Government\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDigital Economy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\u0026nbsp;\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDigital Society\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\u0026nbsp;\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDigital Competence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\u0026nbsp;\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e●\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eConsistency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.931\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.956\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.945\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003ePRI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.915\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.946\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.931\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eCoverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.616\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.606\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.666\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eUnique Coverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIntergroup Consistency Adjustment Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.174\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.139\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.148\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIntragroup Consistency Adjustment Distance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.212\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 133px;\"\u003e\n \u003cp\u003e0.121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 136px;\"\u003e\n \u003cp\u003e0.092\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eOverall Consistency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 387px;\"\u003e\n \u003cp\u003e0.916\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eOverall PRI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 387px;\"\u003e\n \u003cp\u003e0.896\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eOverall Coverage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 387px;\"\u003e\n \u003cp\u003e0.697\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eNote:● indicates the presence of a core condition. All conditions shown are core conditions for their respective configurations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.1.3 Robustness Test\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eRobustness tests were conducted on the digital ecosystem configurations yielding high public service structural optimization:\u0026nbsp;①\u0026nbsp;Increasing the consistency threshold from 0.85 to 0.9;\u0026nbsp;②\u0026nbsp;Raising the case frequency from 2 to 3;\u0026nbsp;③\u0026nbsp;Elevating the PRI from 0.7 to 0.75. The two configurations generated through these three approaches were essentially consistent with the two solutions in the existing configurations. Therefore, the results can be considered robust.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2 Analysis of the Direct Effects of Artificial Intelligence on Public Service Structural Optimization\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.2.1 Descriptive Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe mediating variables in this section are the three digital ecosystem configurations that yielded high public service structure optimization in the sufficiency analysis. By calculating the membership degree of each configuration group, new mediating variable values were derived and subjected to independent regression analysis. The specific measurement methods and descriptive analysis for each variable are presented in Table 4.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 4 Descriptive Statistics\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eVariable Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eVariable Name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMeasurement Description\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObserved Value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStandard Deviation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMinimum\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMaximum\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Independent Variable\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDigital Patent Index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.368\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.087\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.196\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Mediating Variable\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConfiguration 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"3\" valign=\"top\"\u003e\n \u003cp\u003eSet membership degree of each city in the corresponding digital ecosystem configuration\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.329\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.385\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConfiguration 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.315\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.377\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConfiguration 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.386\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.999\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Dependent Variable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePublic Service Structure Optimization\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComposite Index for Public Service Structure Optimization\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.255\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;Control Variables\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGovernment Involvement\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLocal Public Budget Expenditure to GDP Ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.247\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.101\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.107\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.643\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUrbanization Rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eProportion of urban population relative to total population\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.653\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.096\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIndustrial Structure Upgrading\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eValue Added of the Tertiary Industry as a Percentage of GDP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.584\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.751\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.69\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFinancial Level\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTotal Deposits and Loans of Financial Institutions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.785\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.088\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8.16\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003e4.2.2 Regression Analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe regression results are shown in Table 5. Column (1) and Column (2) present the regression results for AI\u0026apos;s impact on optimizing public service structures with and without the inclusion of control variables, respectively. Analysis indicates that regardless of whether control variables are included, the coefficients remain significantly positive, supporting the validity of Hypothesis H1. Given that subsequent regression analysis employs configuration membership values as the mediating variable, the initial regression on raw digital ecosystem data aims to examine whether AI development holistically promotes the digital ecosystem at a macro level. Columns (3) and (4) present regression results for AI\u0026apos;s impact on the digital ecosystem without and with control variables, respectively. Analysis indicates that AI exerts a significant positive influence on the digital ecosystem.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 5 Regression Analysis Results\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"554\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e（1）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e（2）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e（3）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e（4）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eArtificial Intelligence\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.283***\u003c/p\u003e\n \u003cp\u003e(-17.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.146***\u003c/p\u003e\n \u003cp\u003e(-7.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.692***\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(-23.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.392***\u003c/p\u003e\n \u003cp\u003e（-11.34）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eGovernment Involvement\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e-0.219***\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(-4.53) \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e-0.832***\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(-9.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eUrbanization Level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.180**\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e（-3.27）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.0211 \u0026nbsp;\u003c/p\u003e\n \u003cp\u003e（-0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eIndustrial Structure Upgrading\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e-0.0148**\u003c/p\u003e\n \u003cp\u003e(-2.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.0238* \u0026nbsp;\u003c/p\u003e\n \u003cp\u003e（-2.54）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eFinancial Development\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.0319*** \u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(-5.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.0311**\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(-3.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eFixed Effects\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eConstant Term\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.283***\u003c/p\u003e\n \u003cp\u003e（-45.14）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.164***\u003c/p\u003e\n \u003cp\u003e(-5.38)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.270***\u003c/p\u003e\n \u003cp\u003e(-23.88)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.397***\u003c/p\u003e\n \u003cp\u003e(-7.45)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eR2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.6806\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e0.8199\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.7949\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e0.8913\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003eObservations (N)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 103px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eNote：* p\u0026lt;0.05， ** p\u0026lt;0.01，*** p\u0026lt;0.001.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.2.3 Robustness Tests\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTo ensure the robustness of regression results, robustness tests were conducted in four aspects: First, lagging by one period. Given that AI technology applications require time validation, the explanatory variables were lagged by one period and regressed again. Results are shown in Column (1) of Table 6. Second, replacing explanatory variables. Regression was performed using AI word frequency, number of AI enterprises, number of AI industry employees, and industry average salary entropy. Results are shown in Column (2) of Table 6. Third, tail trimming. To mitigate the interference of extreme values on regression outcomes, core variables underwent two-tailed trimming at the 1% level. Results are presented in Column (3) of Table 6. Fourth, we added control variables. We conducted regression analyses incorporating foreign investment levels and total factor productivity, with results shown in Column (4) of Table 6. Fifth, we implemented two-way fixed effects，further controlling for provincial fixed effects in the regression. The above test results indicate that all coefficients for artificial intelligence are positively significant，confirming the robust validity of the regression results.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 6 Robustness Test Results\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"529\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e（1）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e（2）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e（3）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e（4）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e（5）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eArtificial Intelligence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.157***\u003c/p\u003e\n \u003cp\u003e（-8.33）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.277***\u003c/p\u003e\n \u003cp\u003e（-8.01）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.147***\u003c/p\u003e\n \u003cp\u003e（-7.52）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.122***\u003c/p\u003e\n \u003cp\u003e（-5.56）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.0375*\u003c/p\u003e\n \u003cp\u003e（-2.12）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eControl Variables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eFixed Effects\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eConstant Term\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.141***\u003c/p\u003e\n \u003cp\u003e(-4.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.182***\u003c/p\u003e\n \u003cp\u003e（-6.14）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.165***\u003c/p\u003e\n \u003cp\u003e（-5.49）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.150***\u003c/p\u003e\n \u003cp\u003e（-1.47）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-0.0293\u003c/p\u003e\n \u003cp\u003e(-0.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eR2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.8580\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.8283\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.8224\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.8254\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.9800\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003eObservations (N)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e120\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003e4.3 Analysis of Complex Mediation Mechanisms\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e4.3.1 Mediating Effects\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis study further analyzes the impacts and mechanisms of three digital ecosystem configurations on the optimization of public service structures through artificial intelligence, based on the mediation model, to identify indirect effects. The results are shown in Table 7.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(1) Mechanism of the Mature Ecosystem Application. As shown in columns (1) and (2) of Table 7, after introducing the path of the mature ecosystem (Configuration 1), the coefficient of AI decreased from 0.146 to 0.107, while the coefficient of M1 was 0.0494, passing the 1% significance level test. Specifically, the synergistic development of four key elements\u0026mdash;digital infrastructure, digital government, digital society, and digital economy\u0026mdash;forms a virtuous cycle within the digital ecosystem. AI leverages advanced infrastructure to enable data flow, promotes cross-departmental coordination through high-level government mechanisms, and gains sustained momentum from an active digital economy. This pathway not only reduces friction in AI implementation but also continuously enhances the quality and efficiency of public services through internal systemic feedback loops.\u003c/p\u003e\n\u003cp\u003e(2) Administration-Driven Application Mechanism. As shown in columns (1) and (3) of Table 7, the coefficient for the administration-driven pathway (Configuration 2) reached 0.0561 and was statistically significant at the 1% level. After introducing M2, the direct effect of AI further decreased to 0.103. The core advantage of this pathway lies in government-led precision efforts, indicating that even in environments with relatively weak digitalization or industrial foundations, the government can rapidly overcome constraints in key areas through robust policy guidance, resource allocation, and performance incentives. This approach clears institutional barriers for AI adoption and creates favorable implementation conditions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(3) Social Empowerment Application Mechanism. As shown in columns (1) and (4) of Table 7, the coefficient for the social empowerment pathway (Configuration 3) is 0.0656, while the direct effect of AI decreases to 0.0864. This indicates that in regions with relatively limited contributions to the digital economy but active social participation, AI can still leverage inclusive digital infrastructure, widespread social applications, and public digital literacy to form a virtuous cycle of \u0026quot;demand-driven service-responsive.\u0026quot; This pathway does not rely on high-intensity economic investment but instead adopts a socially demand-oriented approach. By enhancing accessibility, inclusiveness, and participation in public services, it achieves more sustainable structural optimization, offering a replicable path for developing regions.\u003c/p\u003e\n\u003cp\u003eOverall, the three pathways exhibit distinct complementarity: the ecologically mature model represents holistic synergy under optimal system conditions; the administratively driven model embodies targeted breakthroughs under government leadership; while the socially empowered model demonstrates inclusive development propelled by societal forces. Together, they constitute a multifaceted implementation mechanism for AI-empowered public services, offering practically viable pathway choices for regions at different developmental stages and with varying resource endowments. Therefore, we accept H2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 7 Complex Mediating Outcomes\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"555\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e（1）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e（2）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e（3）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e（4）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eConfiguration 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e0.0494*** (-3.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eConfiguration 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.0561***\u003c/p\u003e\n \u003cp\u003e(-4.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eConfiguration 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.0656***\u003c/p\u003e\n \u003cp\u003e(-4.78)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eArtificial Intelligence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.146***\u003c/p\u003e\n \u003cp\u003e(-7.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e\u0026nbsp;0.107***\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(-5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.103***\u003c/p\u003e\n \u003cp\u003e(-4.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.0864***\u003c/p\u003e\n \u003cp\u003e(-3.88)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eControl Variables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003eControlled\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eFixed Effects\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eConstant Term\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.164***\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(-5.38)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e0.172***\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e(-5.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.178***\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;(-6.13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.155***\u003c/p\u003e\n \u003cp\u003e(-5.46)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eR2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.8199\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e0.8357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.8382\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e0.8442\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eObservations (N)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 119px;\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003e4.3.2 Bootstrap Robustness Test\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis study employs Bootstrap resampling with 5000 iterations for testing. As shown in Table 8, the confidence intervals for the mediating effects across all three configurations do not include zero, indicating statistically significant results. This confirms that the digital ecosystem configuration exerts a stable mediating effect in the process of AI influencing the optimization of public service structures.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSpecifically, the strength of the mediating effects across the three pathways shows distinct differences. First, the mediating effect accounts for 26.5% in Configuration 1 (Ecosystem Maturity Type), indicating that approximately one-quarter of AI\u0026apos;s impact on optimizing public service structures is achieved through a highly synergistic digital ecosystem. Only when all elements of the digital ecosystem reach a certain level of maturity can its systemic synergistic effects be fully unleashed. Second, the mediating effect in Configuration 2 (Administrative-Driven Type) is stronger, accounting for 29.6%. This indicates that at the current stage of development, the government\u0026apos;s leading role through top-level design, resource allocation, and policy constraints serves as a critical transmission pathway in the AI-driven optimization of public service structures. Compared to the ecosystem maturity model, this pathway does not require waiting for all elements to mature naturally. Instead, it leverages administrative power for selective breakthroughs in key areas, thereby demonstrating higher efficiency in accelerating the rapid implementation of AI applications. Finally, Configuration 3 (Social Empowerment Type) exhibits the most pronounced mediating effect, accounting for 40.9%. This result confirms that broad societal participation and enhanced public digital literacy serve as a critical lever. When citizens possess high digital capabilities and strong engagement awareness, they generate powerful demand pull and oversight functions. This drives public service departments to improve quality and efficiency while simultaneously providing rich application scenarios and data foundations for AI deployment.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFurthermore, the study reaffirms that AI\u0026apos;s direct effect on optimizing public service structures remains significant, though its coefficient exhibits a decreasing trend: from 0.1075 in Configuration 1 to 0.0864 in Configuration 3. This shift indicates that as the mediating mechanism within the digital ecosystem strengthens, AI\u0026apos;s role in optimizing public service structures is increasingly realized through positive feedback loops within the system, thereby reducing the pressure for direct intervention. In summary, a mature, collaborative, or active digital ecosystem can more effectively transmit and amplify the technological dividends of AI.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 8 Bootstrap Robustness Test\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003eMediating Variable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eEffect Type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003eCoefficient\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003eStandard Deviation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 157px;\"\u003e\n \u003cp\u003e95% Confidence Interval\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003eMediation Proportion\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003eConfiguration 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eIndirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0387762\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0..0131539\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.012995\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0645574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e26.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eDirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.1074764\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0211699\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0659841\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.1489687\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003eConfiguration 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eIndirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0432523\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0135792\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0166375\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0698671\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e29.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eDirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.1030003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0212285\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0613931\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.1446074\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003eConfiguration 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eIndirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0598909\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0174463\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0256968\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.094085\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e40.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eDirect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.0863617\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.02304\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.041204\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 79px;\"\u003e\n \u003cp\u003e0.1315193\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"},{"header":"Conclusions and Countermeasures","content":"\u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Research Conclusions\u003c/h2\u003e \u003cp\u003eThis study employs dynamic QCA and regression analysis, examining 30 provinces in China (excluding Tibet) from 2019 to 2023. Integrating digital ecosystem theory and configurational perspectives, it systematically investigates the complex mechanisms through which artificial intelligence drives structural optimization in public services. Notably, it reveals the critical mediating role of digital ecosystems in this process, leading to the following conclusions:\u003c/p\u003e \u003cp\u003e(1) The configuration analysis results indicate that the digital ecosystem plays a crucial role as a multifaceted intermediary. This study innovatively employs three pathways derived from dynamic QCA: the ecologically mature type, the administratively driven type, and the socially empowered type. These pathways are not mutually exclusive but instead exhibit a progressive, convergent, and diffusive evolutionary trend toward optimizing public service structures. Specifically, different Chinese provinces can select distinct development paths based on their developmental stage, resource endowments, and core strengths. For instance: Eastern coastal provinces may prioritize the Ecologically Mature Path, leveraging mature subsystems to pursue higher levels of digital ecosystem synergy; Certain municipalities can effectively employ the Administratively Driven Path to achieve leapfrog development in key industries like information technology and biopharmaceuticals; Central and western regions, with their vast social potential, can deepen the social empowerment path to leverage public strengths and social capital, charting a \"people-centered\" route for structural optimization.\u003c/p\u003e \u003cp\u003e(2) The findings from the complex intermediary mechanism test reveal dual pathways for AI-enabled structural optimization of public services: on one hand, AI promotes the optimization of public service structures; on the other, it drives such optimization at different levels through three distinct types of digital ecosystem intermediaries\u0026mdash;ecologically mature, administratively driven, and socially empowered. This underscores that advancing public service modernization requires not only prioritizing AI technology R\u0026amp;D and application but also cultivating a complementary digital ecosystem. This fosters a virtuous cycle of technological empowerment and systemic optimization, ultimately maximizing public value.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003e5.2 Recommendations\u003c/h2\u003e \u003cp\u003e(1) Implement differentiated and targeted AI development strategies. The three development pathways identified in this study\u0026mdash;ecologically mature, administratively driven, and socially empowering\u0026mdash;offer diverse strategic options for regions at different stages of development. Regions should avoid homogeneous competition through a \"one-size-fits-all\" approach. Instead, they should leverage their unique resource endowments and development priorities to select the most suitable entry points and implement differentiated, targeted AI development strategies. First, for developed regions like Beijing and Shanghai with robust digital infrastructure and vibrant markets, efforts should focus on building open, collaborative innovation platforms. This involves facilitating orderly data flow and value realization while encouraging multi-stakeholder participation\u0026mdash;including enterprises and research institutions\u0026mdash;in scenario-based innovation. The goal is to cultivate a self-evolving, dynamic ecosystem. Second, municipalities like Tianjin and Chongqing, characterized by strong administrative leadership and efficient resource allocation, should concentrate on dismantling \"data silos.\" Through robust top-level design and cross-departmental coordination, they should advance data integration and business process reengineering. Prioritizing breakthroughs in vertical domains such as government services and urban management can drive the entire digital ecosystem's development by significantly enhancing administrative efficiency. Finally, for regions like Henan and Hebei with diverse social needs and active grassroots innovation, policies should moderately tilt downward to encourage community involvement; social organizations; citizen participation in co-governance. By establishing low-barrier tool platforms and incentive mechanisms, AI technology can be leveraged to address pain points in elderly care, education, healthcare, and other public welfare sectors, thereby stimulating the vitality和 resilience of public services from the bottom up.\u003c/p\u003e \u003cp\u003e(2) Refine a multi-stakeholder collaborative digital ecosystem governance system. Deep AI application in public services requires joint participation from government, market, and societal forces. Society-wide efforts should actively build a collaborative mechanism integrating government, enterprises, academia, research, and end-users. Guided by government regulations and targeting enterprise digital-intelligence development, proactive exploration of partnerships with universities and research institutions is essential to effectively align digital technology supply with societal demand. Additionally, governments should vigorously advance the centralized and green deployment of computing power alongside the construction of data resource systems, fully leveraging existing social infrastructure resources to avoid redundant investments in funds and projects. Concurrently, regulatory authorities must accelerate the refinement of AI-related laws and standards\u0026mdash;particularly establishing clear rules regarding data security, algorithmic transparency, and liability determination\u0026mdash;to provide robust safeguards for the dual-engine drive of technological and institutional innovation in public services.\u003c/p\u003e \u003cp\u003e(3) Strengthen information risk prevention and control while building a digital value-driven system. While vigorously promoting AI applications, we must maintain a clear awareness of risks and reinforce a human-centered value orientation. First, establish an evaluation system centered on public satisfaction and sense of fulfillment, eliminating formalism driven by \"technology for technology's sake.\" Second, prioritize the inclusivity and accessibility of digital technology applications. Governments should implement measures such as age-friendly renovations, information accessibility initiatives, and preserving traditional service channels to ensure equal and seamless access to intelligent services for all groups, including the elderly and persons with disabilities. Finally, establish dynamic evaluation and corrective mechanisms to ensure AI development and application consistently align with social ethics and public interest, achieving the organic integration of technological empowerment and humanistic care.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec32\" class=\"Section2\"\u003e \u003ch2\u003e5.3 Limitations and Future Directions\u003c/h2\u003e \u003cp\u003eAlthough this study identifies diverse pathways for AI-empowered public service structural optimization through configuration analysis and confirms the critical mediating role of the digital ecosystem, certain limitations persist, suggesting avenues for future research. First, the current study focuses exclusively on the complex interaction mechanisms between AI and public service structural optimization at the provincial level from 2019 to 2023. Extending the temporal scope to include data from the past decade would enable a more comprehensive understanding of AI's evolutionary trajectory and long-term impacts. Second, future studies could disaggregate AI technologies\u0026mdash;such as generative AI and computer vision\u0026mdash;to systematically examine their distinct operational mechanisms and comparative effectiveness across various public service domains, thereby enhancing policy relevance and technological specificity.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eA Data Availability Statement：\u003c/p\u003e\n\u003cp\u003eThe datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003eCRediT authorship contribution statement:\u003c/p\u003e\n\u003cp\u003ea Chunlin Xiong: Writing-review and editing, Funding acquisition, Supervision\u003c/p\u003e\n\u003cp\u003eb*Ren Fan: Writing \u0026ndash; original draft, Writing \u0026ndash; review and editing, Methodology, Software, Validation, Formal analysis\u003c/p\u003e\n\u003cp\u003ec Yi Liu: Investigation\u003c/p\u003e\n\u003cp\u003eCompeting interests\u003cbr\u003e\u0026nbsp;The author(s) declare no competing interests.\u003c/p\u003e\n\u003cp\u003eFunding:\u003c/p\u003e\n\u003cp\u003eThis research was supported by the Key Project of the National Social Science Fund \u0026quot;Research on the Mechanism of Generating Effectiveness in Rural Digital Governance and its Enhancement Path\u0026quot;(24AZZ010);National Social Science Foundation\u0026rsquo;s Post-funded Project \u0026quot;Evaluation and Optimization of Agricultural and Rural Informatization Policies\u0026quot;(22FGLB006); Hunan Province Social Science Foundation\u0026rsquo;s \u0026ldquo;Academic Hunan\u0026rdquo; High-quality Cultivation Project \u0026quot;Mechanism Innovation for Improving Rural Grassroots Governance Efficiency in the Era of Big Data\u0026quot;(23ZDAJ010); The Key Project of Changsha Soft Science Research Program: Research on Dynamic Evaluation, Simulation Prediction, and Optimi-zation Path of High Quality Development of Rural Digital Economy in Changsha City (KH2502003); Hunan Agricultural University Scientific Research Project \u0026quot;Research on the Driving Mechanism and Optimization Paths of Rural Digital Governance\u0026quot; (25SK015);Hunan Agricultural University Scientific Research Project \u0026quot;Theoretical and Practical Innovations in Digital Rural Development\u0026quot; (25SK027).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMikhaylov, Slava Jankin, Marc Esteve, and Averill Campion. \u0026quot;Artificial intelligence for the public sector: opportunities and challenges of cross-sector collaboration.\u0026quot; Philosophical transactions of the royal society a: mathematical, physical and engineering sciences 376.2128 (2018): 20170357.\u003c/li\u003e\n\u003cli\u003eLatupeirissa, J. J. P., Dewi, N. L. Y., Prayana, I. K. R., Srikandi, M. B., Ramadiansyah, S. A., \u0026amp; Pramana, I. B. G. A. Y. (2024). Transforming public service delivery: A comprehensive review of digitization initiatives. Sustainability, 16(7), 2818.\u003c/li\u003e\n\u003cli\u003eKuziemski, Maciej, and Gianluca Misuraca. \u0026quot;AI governance in the public sector: Three tales from the frontiers of automated decision-making in democratic settings.\u0026quot; Telecommunications policy 44.6 (2020): 101976.\u003c/li\u003e\n\u003cli\u003eCao, H.J., \u0026amp; Xiong, Z.Q. (2025). Co-creation of value in digital public services: Evolutionary origins, theoretical framework, and operational mechanisms. Theoretical Exploration, (03), 56-66.\u003c/li\u003e\n\u003cli\u003eKim, Younhee, Seunghwan Myeong, and Michael J. Ahn. \u0026quot;Living labs for AI-enabled public services: functional determinants, user satisfaction, and continued use.\u0026quot; Sustainability 15.11 (2023): 8672.\u003c/li\u003e\n\u003cli\u003eChang, Jiang, Su, Xin, and Shen, Feiwei. \u0026quot;Exploring New Models of Generative AI-Empowered Public Services.\u0026quot; Scientific Decision Making, 2025, (10): 231-244.\u003c/li\u003e\n\u003cli\u003eHe, Yuanlang, and Yuan, Jianhong. \u0026quot;AI Development and Enhancement of New Quality Productivity: Theoretical Mechanisms and Empirical Tests.\u0026quot; Science and Technology Progress and Policy, 2025, 42(11): 1-11.\u003c/li\u003e\n\u003cli\u003eWirtz, Bernd W., Jan C. Weyerer, and Carolin Geyer. \u0026quot;Artificial intelligence and the public sector\u0026mdash;applications and challenges.\u0026quot; International Journal of Public Administration 42.7 (2019): 596-615.\u003c/li\u003e\n\u003cli\u003eMergel, Ines. \u0026quot;Agile innovation management in government: A research agenda.\u0026quot; Government Information Quarterly 33.3 (2016): 516-523.\u003c/li\u003e\n\u003cli\u003eTan Haibo, Ye Wei. Digital Technology Empowering Public Services: Internal Mechanisms and Primary Approaches [J]. Administrative Forum, 2024, 31(02): 103-110.\u003c/li\u003e\n\u003cli\u003eMargetts, Helen, and Cosmina Dorobantu. \u0026quot;Rethink government with AI.\u0026quot; Nature 568.7751 (2019): 163-165.\u003c/li\u003e\n\u003cli\u003e\u0026Oslash;lnes S, Ubacht J, Janssen M. Blockchain in government: Benefits and implications of distributed ledger technology for information sharing[J]. Government information quarterly, 2017, 34(3): 355-364. \u003c/li\u003e\n\u003cli\u003eBurlacu S, Dobre F, Mateescu V M, et al. From bureaucracy to e-Government: Digital transformation of Public Administration in Romania for a sustainable future[J]. European Journal of Sustainable Development, 2025, 14(3): 289-289.\u003c/li\u003e\n\u003cli\u003eJanowski, Tomasz. \u0026quot;Digital government evolution: From transformation to contextualization.\u0026quot; Government information quarterly 32.3 (2015): 221-236.\u003c/li\u003e\n\u003cli\u003eYao Jiaquan, Zhang Kunpeng, Guo Lipeng, et al. How Does Artificial Intelligence Enhance Corporate Productivity? \u0026mdash; A Perspective Based on Labor Skill Structure Adjustment [J].Management World, 2024, 40(02): 101-116+133+117-122.\u003c/li\u003e\n\u003cli\u003eFarrell, Eimear, et al. \u0026quot;Artificial intelligence for the Public Sector.\u0026quot; (2023).\u003c/li\u003e\n\u003cli\u003eGarcia Alonso, Roberto, and Sebastian Lippez-De Castro. \u0026quot;Technology helps, people make: A smart city governance framework grounded in deliberative democracy.\u0026quot; Smarter as the new urban agenda: A comprehensive view of the 21st century city. Cham: Springer International Publishing, 2015. 333-347.\u003c/li\u003e\n\u003cli\u003eBurlacu S, Dobre F, Mateescu V M, et al. From bureaucracy to e-Government: Digital transformation of Public Administration in Romania for a sustainable future[J]. European Journal of Sustainable Development, 2025, 14(3): 289-289.\u003c/li\u003e\n\u003cli\u003eLiu Haimin, Wu Hailong. \u0026quot;Problem Patterns, Multiple Causes, and Countermeasures of Digital Formalism in Educational Governance.\u0026quot; Modern Education Management, 1-14 [2025-10-24].\u003c/li\u003e\n\u003cli\u003eLinders, Dennis. \u0026quot;From e-government to we-government: Defining a typology for citizen coproduction in the age of social media.\u0026quot; Government information quarterly 29.4 (2012): 446-454.\u003c/li\u003e\n\u003cli\u003eRadanliev, Petar, et al. \u0026quot;Ethics and responsible AI deployment.\u0026quot; Frontiers in Artificial Intelligence 7 (2024): 1377011.\u003c/li\u003e\n\u003cli\u003eGarcia-Castro, Roberto, and Miguel A. Ari\u0026ntilde;o. \u0026quot;A general approach to panel data set-theoretic research.\u0026quot; Journal of Advances in Management Sciences \u0026amp; Information Systems 2.63-76 (2016): 526.\u003c/li\u003e\n\u003cli\u003eLi Deguo and Chen Yiqing. \u0026quot;Risks and Prevention of Artificial Intelligence-Powered Public Services from the Perspective of \u0026apos;Normal Accidents\u0026apos;.\u0026quot; *Social Sciences in Chinese Universities*, 2025, (05): 145-155+159.\u003c/li\u003e\n\u003cli\u003eLi Liqing, Luan Wenting. Generative Artificial Intelligence in Public Service Governance: Mechanisms of Action and Risk Mitigation Strategies. *Journal of Nanchang University (Humanities and Social Sciences)*, 2025, 56(03): 113\u0026ndash;124.\u003c/li\u003e\n\u003cli\u003eHu Hongbin. The Driving System, Operational Mechanism, and Pathways for Enhancing the Development of an Education Nation in the New Era: A Systemic Analytical Framework. Journal of Southwest Minzu University (Humanities and Social Sciences Edition), 2024, 45(07): 186-195.\u003c/li\u003e\n\u003cli\u003eWu Lingling. On the Systemic Allocation and Optimization Pathways of Library Resources: A Novel Analytical Framework. Library Journal, 2022, (02): 101-106.\u003c/li\u003e\n\u003cli\u003eBusuioc, Madalina. \u0026quot;Accountable artificial intelligence: Holding algorithms to account.\u0026quot; Public administration review 81.5 (2021): 825-836.\u003c/li\u003e\n\u003cli\u003eOsborne S P, Cucciniello M, Nasi G, et al. Digital transformation, artificial intelligence and effective public services: challenges and opportunities[J]. Global Public Policy and Governance, 2022, 2(4): 377-380.\u003c/li\u003e\n\u003cli\u003eWang P. Connecting the parts with the whole: toward an information ecology theory of digital innovation ecosystems[J]. MIS quarterly, 2021, 45(1): 397-422.\u003c/li\u003e\n\u003cli\u003eXiang Meng, Wang Zhigang. Research on Data Element Market Governance Strategies from a Digital Ecology Perspective[J]. Xinjiang Social Sciences, 2023,(06):54-61.\u003c/li\u003e\n\u003cli\u003eWang Jiandong, Dou Yue, Ren Junxia, et al. From \u0026quot;Holistic Services\u0026quot; to \u0026quot;Intelligent Entity Services\u0026quot;: Research on Innovation in Government Public Service Models Under Digital Transformation [J]. E-Government, 2023, (01): 12-22.\u003c/li\u003e\n\u003cli\u003eWang Jing, Zhang Hong. Application of Artificial Intelligence in Digital Government Construction and Its Path to Legalization: Analysis Based on DeepSeek\u0026apos;s Application Scenarios in Digital Administration [J]. Chinese and Foreign Law, 2025, 37(03): 625-643.\u003c/li\u003e\n\u003cli\u003eZhang Hongchun, Yang Huan. A Digital Literacy Framework for Civil Servants in the Context of Digital Government: A Conceptual Model [J]. E-Government, 2023, (01): 110-124.\u003c/li\u003e\n\u003cli\u003eCai Zhenhua, Zhao Youhua. Governance of Public Service Demand in the Era of Artificial Intelligence: Drivers and Directions [J]. Ningxia Social Sciences, 2020, (02): 47-54.\u003c/li\u003e\n\u003cli\u003eBu Yanjun. Analysis of Power Diffusion Trends and Challenge Responses in the Context of Artificial Intelligence Development [J]. Studies in Science of Science, 2025, 43(03): 608-616.\u003c/li\u003e\n\u003cli\u003eXue Xinyi, Zhou Jie. Analysis of the Driving Mechanism and Countermeasures for Promoting High-Quality Development of Public Cultural Services by Generative Artificial Intelligence [J]. Library Theory and Practice, 2025, (05): 14-23.\u003c/li\u003e\n\u003cli\u003eBonezzi, Andrea, and Massimiliano Ostinelli. \u0026quot;Can algorithms legitimize discrimination?.\u0026quot; Journal of Experimental Psychology: Applied 27.2 (2021): 447.\u003c/li\u003e\n\u003cli\u003eDu Yunzhou, Sun Ning, Liu Qiuchen. Developing and Analyzing Complex Mediating Models Using Mixed Methods: The Case of Business Environment Promoting Innovation Vitality, Synergizing New Quality Productivity, and \u0026quot;Employment First\u0026quot; [J]. Management World, 2024, 40(06): 217-237.\u003c/li\u003e\n\u003cli\u003eZhang Liqiong. High-Quality Development of Public Services, Disruptive Innovation, and Common Prosperity in Urban and Rural Areas [J]. Statistics and Decision Making, 2025, 41(11): 95-99.\u003c/li\u003e\n\u003cli\u003eHu Hongbin. The Driving System, Operational Mechanism, and Improvement Pathways for Advancing Education Powerhouse Construction in the New Era: A Systematic Analytical Framework [J]. Journal of Southwest University for Nationalities (Humanities and Social Sciences Edition), 2024, 45(07): 186-195.\u003c/li\u003e\n\u003cli\u003eSun Daming, Hu Sumin, Zhu Tianyi, et al. How Industrial Internet Platforms Drive New Quality Productivity Development: A Data Element Perspective [J]. Science and Technology Progress and Policy, 2025, 42(03): 38-49.\u003c/li\u003e\n\u003cli\u003eShan Wei, Xie Wenbin, Sun Yizhong, et al. Breakthrough Innovation and High-Quality Development of Low-Altitude Economy Enterprises[J/OL]. Studies in Science of Science, 1-26[2025-11-04].\u003c/li\u003e\n\u003cli\u003eHu Hongbin. The Driving System, Operational Mechanism, and Improvement Pathways for Advancing Education Powerhouse Construction in the New Era: A Systematic Analytical Framework [J]. Journal of Southwest University for Nationalities (Humanities and Social Sciences Edition), 2024, 45(07): 186-195.\u003c/li\u003e\n\u003cli\u003eGarcia-Castro, Roberto, and Miguel A. Ari\u0026ntilde;o. \u0026quot;A general approach to panel data set-theoretic research.\u0026quot; Journal of Advances in Management Sciences \u0026amp; Information Systems 2.63-76 (2016): 526.\u003c/li\u003e\n\u003cli\u003eZhang, Jichang, Long, Jing, and Wang, Zemin. \u0026quot;What Kind of Institutional Environment Promotes High Entrepreneurial Activity? A Dynamic QCA Analysis Based on Provincial Panel Data.\u0026quot; Science and Technology Progress and Policy, 2024, 41(24): 36-48.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial Intelligence, Optimization of Public Service Structure, Digital Ecosystem, Complex Mediation, Dynamic QCA","lastPublishedDoi":"10.21203/rs.3.rs-8350814/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8350814/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAgainst the backdrop of the \"Digital China\" strategy, the digital ecosystem has emerged as a strategic fulcrum for transforming public service delivery from a supply-oriented to a demand-oriented paradigm, with artificial intelligence (AI) serving as the core engine of structural optimization of public services. However, existing research has rarely examined the complex causal mechanisms by which AI, through configurations of digital ecosystem elements, shapes the structural optimization of public services. Based on panel data from 30 Chinese provinces covering 2019–2023, this study integrates the configurational perspective with digital ecosystem theory to develop a systematic“technology input–system integration–value output”framework. Employing an advanced mediation model that combines dynamic qualitative comparative analysis (QCA) and regression methods, the research elucidates the pathways and process mechanisms by which AI drives the structural optimization of public services. The findings indicate that AI significantly promotes the structural optimization of public services, achieving multi-level upgrades through three distinct configurations of the digital ecosystem: (1) ecologically mature type (an advanced, all-encompassing development model), (2) administratively driven type (a government-led approach), and (3) socially empowered type (an inclusive development pathway).This study clarifies the mechanism by which AI and the digital ecosystem collaboratively reshape the architecture of public services, highlighting that achieving public service modernization requires not only strengthening he development and application of AI technologies but also fostering a compatible digital ecosystem. Such synergy creates a virtuous cycle of technological empowerment and systemic optimization, ultimately maximizing public value. The results provide empirical evidence and forward-looking insights for inclusive, digitally integrated governance in the era of intelligent technologies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJEL:\u003c/strong\u003eH83; O33\u003c/p\u003e","manuscriptTitle":"How Does Artificial Intelligence Drive the Optimization of Public Services Structural? ——Complex Intermediary Mechanisms Based on the Digital Ecosystem","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-26 10:44:36","doi":"10.21203/rs.3.rs-8350814/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"017caa9a-2a34-4f0a-80a3-60dea8f41a63","owner":[],"postedDate":"December 26th, 2025","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-05-20T06:59:34+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-14T10:05:02+00:00","index":105,"fulltext":""},{"type":"reviewerAgreed","content":"286910930915890846969755933360615749625","date":"2026-05-14T09:58:33+00:00","index":104,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":60202954,"name":"Business and commerce/Information systems and information technology"},{"id":60202955,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-05-20T07:11:01+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-26 10:44:36","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8350814","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8350814","identity":"rs-8350814","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}