Technological Transformation of Environmental Governance: Evidence from a Quasi-Natural Experiment in China

Technological Transformation of Environmental Governance: Evidence from a Quasi-Natural Experiment in China | 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 Technological Transformation of Environmental Governance: Evidence from a Quasi-Natural Experiment in China Zhiwei Liu, Haiming Liu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7322815/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 As environmental governance becomes a central issue of global concern, how to use emerging technologies to improve government environmental performance has become an important area of research. This study examines the impact of the establishment of China’s Artificial Intelligence (AI) Innovation Pilot Zones on local government environmental performance, with a focus on the key role of government in environmental governance. Based on data from three batches of AI innovation pilot zones between 2019 and 2021, we find that AI pilot zones significantly enhance local government environmental performance by promoting green innovation and strengthening environmental regulations. However, higher fiscal autonomy to some extent weakens the actual effectiveness of environmental policies, while leaders with higher education levels are more effective in implementing environmental policies. This study expands the understanding of technology-driven environmental governance and provides new theoretical perspectives for environmental policy in emerging economies, offering empirical support for the role of technology in empowering public governance. Earth and environmental sciences/Environmental sciences Earth and environmental sciences/Environmental social sciences Social science/Environmental studies Artificial Intelligence Government Environmental Performance Fiscal Autonomy Leader Education Level Figures Figure 1 Figure 2 1. Introduction In recent years, environmental issues have emerged as a central focus of global concern. In this regard, China is at the forefront due to its severe pollution, which poses a substantial challenge to the sustainable development of both its economy and society. While China has achieved remarkable economic growth over the past few decades, the accompanying environmental pressures have intensified (Chen & Fleisher, 1996 ; Li & Lin, 2017 ). In 2010, more than two-thirds of the world’s 50 most polluted cities were in China, and the country’s pollution emissions ranked among the highest globally. Despite the Chinese government’s proactive policy initiatives to address environmental degradation, local governments continue to face significant challenges in policy implementation, particularly in relation to conflicts of interest and information asymmetry, which substantially undermine the effectiveness of environmental governance (Cao et al., 2023 ; Greenstone et al., 2022 ). In this context, emerging technologies, particularly artificial intelligence (AI), are increasingly viewed as promising tools for alleviating information asymmetries and enhancing the effectiveness of environmental governance. However, whether these technologies can effectively address the principal-agent problems inherent in environmental governance remains an unresolved and complex issue, warranting further scholarly exploration. AI, a key driver of technological and industrial transformation, has the potential to significantly enhance productivity, foster innovation, and stimulate economic growth (Acemoglu & Restrepo, 2018 ; Li et al., 2025 ; W. J. Liu et al., 2024 ; Yao et al., 2024 ). Existing research has predominantly focused on the impact of AI on corporate environmental performance, with numerous studies demonstrating the crucial role of AI technologies in improving the environmental outcomes of firms (Ying et al., 2023 ; Zhang et al., 2025 ; Zhou et al., 2024 ). On one hand, AI directly contributes to environmental performance by optimizing production processes, improving resource efficiency, and promoting green innovation. AI not only enhances the quantity of corporate green innovations, but also increases their innovation efficiency (Dong et al., 2025 ; Hussain et al., 2024 ; Z. Wang et al., 2024 ). Intelligent manufacturing and energy distribution systems allow AI to significantly reduce energy waste, thereby optimizing environmental performance (Ghobakhloo & Fathi, 2021 ; Kluczek et al., 2021 ). Furthermore, AI offers distinct advantages in intelligent scheduling and energy demand forecasting, enabling firms to effectively reduce carbon dioxide emissions and providing a technological foundation for achieving green development (Kanabkaew et al., 2019 ; Sun et al., 2022 ). However, this governance model is not without controversy. External regulatory pressures may encourage firms to engage in symbolic environmental actions rather than substantive improvements. While green finance policies have accelerated corporate green transformation, they often come with financial constraints, prompting some high-polluting firms to resort to “greenwashing.” These firms may superficially comply with environmental standards without implementing genuine green innovations (Hu et al., 2023 ; Wang et al., 2023 ; Yi et al., 2025 ; Zhang, 2022 ; Zhang, 2023 ). Government attention and involvement (as reflected in governmental documents and policies) represent the most direct factors influencing environmental quality (Zhang & Wu, 2018 ). In recent years, both domestic and international scholars have increasingly focused on evaluating the effectiveness of government environmental governance. Research indicates that air pollution regulation in developing countries can effectively reduce particulate matter in the air, thereby accelerating the process of improving environmental quality (Greenstone & Hanna, 2014 ; Luo & Li, 2018 ; Yang & Chou, 2018 ). Within the realm of environmental governance, the application of technologies, particularly automated monitoring systems, is crucial for enhancing governance effectiveness and ensuring compliance. For instance, the introduction of automated air pollution monitoring has not only effectively curbed local government manipulation of data, but also significantly improved the accuracy of air quality reports (Greenstone et al., 2022 ; Greenstone et al., 2021 ). The use of such technologies has brought real-time data and transparency, thereby advancing policy implementation and reducing information biases and deficiencies in the governance process. Studies have shown that automated monitoring systems have prompted local governments to respond swiftly to pollution hotspots, substantially reducing pollution levels in monitored areas (Yang et al., 2024 ). Moreover, as governments continue to strengthen pollution control policies, environmentally supportive policies backed by technology have effectively reduced pollutant emissions and improved air quality (Cao et al., 2023 ; Jiang et al., 2021 ). However, despite the pivotal role of technology in enhancing environmental governance, issues such as technological lag and data manipulation persist, necessitating the safeguarding of the fairness and effectiveness of technology application within institutional frameworks (Z. F. Liu et al., 2024 ). Consequently, while AI serves as a key driver of technological and societal transformation, its impact on government environmental performance remains uncertain. Furthermore, insufficient attention is given to the actual occurrence and mechanisms of this technology, which makes it difficult for both scholars and policymakers to accurately assess the real effectiveness of AI in improving government environmental performance. It also complicates comprehensively evaluating the potential implications of such reforms on central-local government relations. Considering this, this study focuses on the pilot cities of the three batches of “Artificial Intelligence Innovation Development Experimental Zones”(AIIDEZ) established in China from 2019 to 2021. Using environmental performance as the key outcome variable, this research systematically examines the strategic choices made by local governments in different contexts following the establishment of these AI innovation zones. Unlike previous studies that focused on corporate-level environmental performance, this research centers on the role of government in driving environmental improvements, positing governmental policy guidance and governance capacity as the critical drivers of environmental enhancement. By constructing a panel dataset of 30 Chinese provinces from 2011 to 2022 and applying multiple robustness checks, we rigorously analyze the impact of AIIDEZ on environmental performance and explore the underlying mechanisms. The empirical results demonstrate that this reform pilot has significantly improved government environmental performance by promoting green innovation and strengthening environmental regulation. Furthermore, the study reveals that government leaders with a higher level of education are more sensitive to and proactive in responding to the pilot reforms. However, when local governments have relatively greater fiscal autonomy, officials tend to prioritize economic development, thereby diminishing the actual effectiveness of the reform pilot. Existing research on government environmental governance and air quality lays the foundation for our study and opens avenues for further exploration. The innovative contributions of our research are as follows: First, it deepens understanding of the consequences of AI application. Previous studies focused on the impact of AI applications on the economy (Yang, 2022 ; Yao et al., 2024 ) and corporate environmental performance (Guo et al., 2025 ; A. Wang et al., 2024 ), often overlooking the crucial role of government as a key factor in the application of AI. By examining the relationship between AI application and government environmental performance, this study expands understanding of AI’s broader implications. Second, this research contributes to the study of environmental performance in emerging economies. While existing literature primarily explores the impact of internal technological advancements on the environmental performance of emerging economies (Jin et al., 2024 ; Liu et al., 2025 ), this study offers a novel perspective by investigating the effect of AI application on government environmental performance through the lens of internal principal-agent incentives, enriching knowledge on environmental performance in emerging economies. Finally, from a qualitative perspective, this study analyzes the “Chinese-style” logic and mechanisms underlying the significant outcomes of AIIDEZ, providing new insights for governments to continue implementing environmental governance policies and achieving harmonious coexistence between humans and nature. 2. AIIDEZ in China Driven by globalization and modernization, government organizations often face multiple competing tasks in A real-world governance context. For instance, when advancing regional economic development with a focus on high-quality growth, governments frequently encounter the dilemma of balancing development and protection. Economic growth typically relies on resource-intensive and high-pollution industries, the expansion of which inevitably leads to environmental pollution, ecological degradation, and a decline in social welfare. Improving environmental performance often necessitates costly technological upgrades or industrial restructuring, placing local governments under pressure to meet short-term economic goals while simultaneously addressing long-term environmental sustainability challenges. In 2003, Chinese President Hu Jintao introduced the “Scientific Outlook on Development,” emphasizing the need to balance economic growth with environmental protection. Since then, promoting sustainable development has been a central component of China’s national strategy. In practice, environmental governance in China follows a primary “command and control” model, which is similar to the target-based system under the planned economy. At the beginning of each year, the central government sets nationwide targets for the reduction of various pollutants. These targets are then allocated to provinces, which then assign them to municipalities and counties. To incentivize compliance, the achievement of environmental governance goals is directly linked to the political advancement of key government leaders (Liang & Langbein, 2015 ). In 2007, the State Council issued “Measures for the Assessment of Total Pollutant Emissions,” implementing a “one strike and you’re out” rule for government leaders who failed to meet their pollution reduction targets. Such leaders were ineligible for top performance evaluations. Furthermore, the central government revised the “Environmental Protection Law of the People's Republic of China” in 2014, officially establishing the legal responsibility of governments at all levels for environmental quality within their jurisdictions, and clearly defining emission standards for various pollutants (Wang, 2022 ). Despite the central government’s long-standing emphasis on environmental governance, in practice, government negligence in environmental regulation is often difficult to detect and correct in a timely manner. To address these issues, the central government has gradually introduced a series of environmental supervision systems, including environmental protection inspections characterized by “enterprise supervision” (referred to as “environmental inspections”) and central environmental protection inspections marked by “government supervision” (referred to as “central environmental inspections”) (Zhu & Wang, 2024 ; Zhuang, 2024 ), However, while these reforms, which focus on “enterprise supervision” and “government supervision,” have succeeded in partially centralizing environmental power, they have not fundamentally altered the issue of the excessive centralization of environmental authority. Some studies have found that although environmental performance improves significantly during inspection periods, the policy effects immediately diminish once the inspections are over, and the government’s enforcement bias persists (Kou et al., 2022 ; Razzaq et al., 2023 ; Tan & Mao, 2021 ; Wu & Hu, 2019 )。 In the face of this complex situation, AI, a disruptive innovation, offers a potential pathway for enhancing environmental governance capabilities. By integrating technologies such as big data analytics, machine learning, and the Internet of Things (IoT), AI can provide efficient technical support across various domains including environmental monitoring, resource optimization, and pollution control. Through AI, local governments can achieve more precise, dynamic, and intelligent environmental governance. Moreover, as a new production factor, AI can improve management and organizational efficiency, facilitating the intelligent and green transformation of economic activities. This enables the transition of China’s traditional industries towards higher-end, smarter, and greener models, thereby alleviating the conflict between economic development and environmental protection. Against this backdrop, in 2019, the Ministry of Science and Technology of China issued the “Guidelines for the Construction of National New Generation AI Innovation Development Experimental Zones.” To date, 18 cities including Beijing have been designated as AIIDEZ, The list is provided in Appendix A. The establishment of these zones aims to promote the application and industrialization of AI technologies through policy support and resource allocation, serving as a key driver for high-quality economic development. What are the policy effects of these zones? Have they significantly improved government environmental performance? What are the underlying mechanisms? Clarifying these questions is of crucial practical significance for innovating AI development models, defining AI’s future trajectory, enhancing environmental performance, and driving high-quality economic development. 3. Empirical design 3.1. Dependent variable The primary dependent variable in this study is government environmental performance. To measure this, we use the PM2.5 data from the China High Air Pollutants (CHAP) dataset, which is published by the National Qinghai-Tibet Plateau Science Data Center (Wei et al., 2020 ; Wei et al., 2021 ). In the robustness checks section (section 4.2 ), we replace the dependent variable with global surface PM2.5 concentration data (µg/m³) from the University of Washington St. Louis campus and AQI data from the National Urban Air Quality Real-time Release Platform of the China National Environmental Monitoring Center. This substitution ensures the robustness of our study. 3.2. Independent variables The primary explanatory variable in this study is a binary policy indicator variable \(\:{Treat}_{i,t}\) 。If province \(\:i\) is approved as an AIIDEZ in year \(\:t\) , the \(\:{Treat}_{i,t}\) for the corresponding province is assigned a value of 1 for year \(\:t\) and subsequent years; otherwise, it is assigned a value of 0. The Ministry of Science and Technology of China established pilot cities in three batches in 2019, 2020, and 2021. Since some pilot cities were approved toward the end of the year, we account for the lag effect of the policy by assigning cities approved after September to the following year. Figure 1 illustrates the timeline of the AIIDEZ across different regions in China. As Fig. 1 shows, reform implementation exhibits notable spatiotemporal variation. The program launched in Beijing and Shanghai in 2019, gradually expanding to other eastern provinces and the central-western regions thereafter. The deep red areas represent regions that implemented AIIDEZ earlier, while the lighter red areas indicate those that adopted the reform later. The gray-shaded areas reflect the limitations of the reform’s coverage. This spatial distribution of AIIDEZ highlights the dynamic progression of AI development in China. 3.3. Control variables Urban socioeconomic data In addition, we control for various urban socioeconomic factors including per capita GDP, public budget revenue, and resident population. The data primarily comes from the China Statistical Yearbook. Urban environmental pollution data The environmental performance of a province may be influenced by the local level of environmental pollution. Therefore, we control for total wastewater discharge, sulfur dioxide emissions from exhaust gases, the generation of general industrial solid waste, and total electricity consumption. The data mainly comes from the China Statistical Yearbook and the China Environmental Protection Statistical Yearbook. Urban meteorological data The environmental performance of a province may also be affected by its geographical factors. Therefore, we control for various covariates related to a province’s climate characteristics including average temperature, annual cumulative precipitation, and average wind speed. These data are processed by the National Centers for Environmental Information (NCEI) under the National Oceanic and Atmospheric Administration (NOAA). Descriptive statistics for the key variables are presented in Table 1 . 3.4. Model specifications This study treats the establishment of AIIDEZ in China as a quasi-natural experiment and rigorously evaluates the impact of these pilot zones on local environmental performance using a difference-in-differences (DID) model. Given that the implementation of AIIDEZ is staggered over time, it is not possible to create a single uniform dummy variable to reflect the establishment time across all regions. Therefore, we classify provinces with established AI innovation zones as the treatment group, and those without as the control group. In addition, we identify the specific start time for each province’s participation in the pilot program to define the policy dummy variable. The empirical model structure reflecting this setup is as follows: $$\:{PM2.5}_{i,t}=\alpha\:+{\beta\:\:Treat}_{i,t}+{\gamma\:X}_{i,t}+{\mu\:}_{i}+{\lambda\:}_{t}+{\epsilon\:}_{i,t}$$ 1 In this model, \(\:i\) denotes the province and \(\:t\) the year. \(\:{PM2.5}_{i,t}\) represents environmental performance. \(\:{Treat}_{i,t}\) is the primary explanatory variable, which indicates whether province \(\:i\) was designated as an AIIDEZ in year \(\:t\) . This study focuses on the coefficient \(\:\beta\:\) of \(\:{Treat}_{i,t}\) , as it reflects the net effect of establishing an AIIDEZ on local environmental performance. \(\:{X}_{i,t}\) includes a range of province-specific factors, covering natural climate variables and socioeconomic factors. In the empirical framework, \(\:{\mu\:}_{i}\) represents province fixed effects, \(\:{\lambda\:}_{t}\) year fixed effects, and \(\:{\epsilon\:}_{i,t}\) the random disturbance term. Table 1 Descriptive statistics of major variables. N Mean SD Min Max Log PM2.5 360 3.641 0.374 2.664 4.535 Treat 360 0.117 0.321 0 1 Log PerGDP 360 10.87 0.461 9.682 12.15 Log Fiscal revenue 360 7.650 0.844 5.023 9.554 Log Population 360 8.208 0.741 6.342 9.448 Log Temperature 360 2.547 0.499 0.890 3.255 Log Precipitation 360 6.849 0.534 4.993 7.843 Log Wind 360 2.185 0.170 1.674 2.661 Log Waste water 360 12.05 0.819 9.966 13.75 Log Solid waste 360 2.984 1.368 -2.207 5.208 Log SO2 360 8.986 1.062 5.142 10.86 Log Electricity consumption 360 7.435 0.696 5.222 8.971 4. Empirical results 4.1. Baseline results Table 2 provides a detailed description of the impact of the establishment of AIIDEZ on environmental performance. All Variables Used in Baseline Result listed in Appendix B. Model 1 estimates the simple binary relationship between the AI innovation development zones and environmental performance. Model 2 includes macroeconomic control variables at the province level, and Model 3 further incorporates meteorological control variables. Model 4, as the comprehensive model, additionally includes environmental pollution data at the province level. All models show that the coefficient of Treat is significantly negative at the 1% significance level. Focusing on Model 4, the estimated values indicate that the establishment of experimental zones has led to an average reduction of 8% in PM2.5 levels, suggesting the effective role of AI innovation zones in reducing PM2.5 . While these findings provide strong evidence, the specific mechanisms through which AIIDEZ achieves these environmental outcomes remain uncertain. Further investigation of these potential pathways will deepen our understanding of the operational dynamics of this policy. Table 2 Baseline results DV: PM2.5 (Log) (1) (2) (3) (4) Treat -0.088*** -0.095*** -0.094*** -0.080*** (0.025) (0.026) (0.024) (0.016) Economic controls N Y Y Y Meteorology controls N N Y Y Environmental pollution controls N N N Y Year FE Y Y Y Y Province FE Y Y Y Y Within R² 0.102 0.132 0.143 0.303 Observation 360 360 360 360 Notes : Standard errors in parentheses are clustered at province level, * p < 0.10, ** p < 0.05, *** p < 0.01. 4.2. Parallel trend and robustness tests The validity of the DID estimation relies on the parallel trend assumption. This assumption posits that the provinces that implemented AIIDEZ exhibit no systematic differences in the PM2.5 concentration change trend compared to the provinces that did not implement the experimental zones before the introduction of the pilot program. To test this assumption, this study follows the event study method introduced by (Jacobson et al., 1993 ) to conduct a parallel trends test on the model. The model is set as follows: $$\:{Y}_{i,t}=\alpha\:+{\sum\:_{k\ge\:-5}^{k}\beta\:\:Treat}_{i,{t}_{0}+k}+{\gamma\:X}_{i,t}+{\mu\:}_{i}+{\lambda\:}_{t}+{\epsilon\:}_{i,t}$$ 2 In this model, the variable \(\:{Treat}_{i,{t}_{0}+k}\) represents the event window dummy variable indicating the periods before and after the implementation of the AIIDEZ policy. \(\:{t}_{0}\) is the year the province completed the establishment of AIIDEZ, and \(\:{t}_{0}+k\) refers to the various years before and after the policy pilot. The value of k takes on values of -5, -4, -3, -2, -1, 0, 1, 2, and 3, with the study introducing dummy variables for the five years prior to the pilot. Any year greater than or equal to five years is grouped into the five-year category. The study uses the year immediately preceding the establishment of AIIDEZ as the reference group, with other control variables and the baseline model remaining similar. Figure 2 presents the parameter estimates and their corresponding confidence intervals. As the figure shows, for periods where \(\:k<0\) , the value of \(\:\beta\:\) does not significantly deviate from zero, indicating that prior to the reform, there were no significant differences in the trends of PM2.5 concentrations between provinces that implemented AIIDEZ and those that did not. This pattern satisfies the parallel trends assumption. In contrast, at \(\:k=0\) , the value of \(\:\beta\:\) deviates notably from zero, showing a sharp decline. In subsequent periods, \(\:\beta\:\) gradually decreases. This suggests that the implementation of AIIDEZ significantly slowed the increase in PM2.5 concentrations in the provinces that adopted the policy. Over time, this effect strengthens. In addition, we conducted a series of robustness checks to address concerns regarding measurement errors and potential omitted variables. These checks included analyzing different subsamples, considering alternative time division standards, accounting for the impact of other environmental policies during the policy period, and using alternative dependent variables. The results remained consistent with the baseline findings. Further details are provided in Appendix C. 4.3. Mechanism analysis As the government’s environmental governance functions evolve, it is necessary to further explore the underlying mechanisms through which PM2.5 concentrations are reduced. To this end, we delve into the potential channels through which AIIDEZ reduce PM2.5 concentrations, focusing on green innovation and the intensity of environmental regulations. Based on (Dell, 2010 ), the model is set as follows: $$\:Med=\alpha\:+{\beta\:\:Treat}_{i,t}+{\gamma\:X}_{i,t}+{\mu\:}_{i}+{\lambda\:}_{t}+{\epsilon\:}_{i,t}$$ 3 Where \(\:Med\) represents the mechanism variables, which are proxied by Green Innovation and the Intensity of Environmental Regulations . The other variables are set in accordance with the baseline model. 4.3.1. Green innovation Theoretically, AIIDEZ can reduce PM2.5 concentrations by promoting green innovation. To explore this potential mechanism, this study uses the number of green patent applications and granted patents from the China National Research Data Service Platform (CNRDS) as proxies for green innovation. The regression analysis results in columns (1)–(4) of Table 3 indicate that AIIDEZ effectively enhances local green innovation, consistent with theoretical expectations. This relationship may be attributed to the fact that AIIDEZ fosters green innovation through policy support, technological innovation, and industrial transformation, thereby reducing PM2.5 concentrations. Policy incentives have accelerated the growth of green patents, AI technology has improved the efficiency of green technologies, and industrial transformation has driven the green upgrading of high-polluting industries. These factors have worked synergistically to improve environmental quality. In summary, the study confirms the key role of green innovation in effectively reducing environmental pollution, further emphasizing the important mechanism through which AIIDEZs enhance air quality by promoting green innovation. 4.3.2. Environmental regulation intensity Existing research has shown that a key factor contributing to environmental pollution is the government’s tendency to lower environmental regulations in pursuit of economic benefits, resulting in “local regulatory bias” (Jia & Nie, 2017 ). With the establishment of AIIDEZ, the government's environmental regulatory intensity and resource allocation in environmental governance may change. Given that the government’s environmental regulatory intensity is to some extent a reflection of the costs associated with pollution control, the stronger the environmental regulations, the higher the government’s pollution control costs. Therefore, we measure environmental regulation by the ratio of government industrial pollution control investment to industrial added value. The regression results in columns (5) and (6) of Table 3 show that AIIDEZ significantly enhances the government’s environmental regulatory intensity. This suggests that the government may optimize environmental governance by enhancing regulatory precision, improving decision-making efficiency, and reducing governance costs. With the help of AI technology, the government can achieve accurate pollution source identification and real-time monitoring, thereby improving the scientific and transparent nature of governance, and reducing information asymmetry and regulatory blind spots in traditional governance models. This shift alleviates the problem of local governments’ “regulatory bias” and provides a new path for promoting more efficient and sustainable environmental governance. Table 3 The mechanism of AIIDPZ alleviates local PM2.5. DV: Green Innovation DV: Environmental Regulation Intensity (1) (2) (3) (4) (5) (6) Treat 6482.485*** 5252.981*** 6147.204*** 5687.539*** 0.002*** 0.002*** (1734.179) (1611.412) (1726.790) (1657.605) (0.001) (0.001) Economic controls N Y N Y N Y Meteorology controls N Y N Y N Y Environmental pollution controls N Y N Y N Y Year FE Y Y Y Y Y Y Province FE Y Y Y Y Y Y Within R² 0.120 0.326 0.178 0.325 0.042 0.122 Observation 360 360 360 360 360 360 Notes : Standard errors in parentheses are clustered at province level, * p < 0.10, ** p < 0.05, *** p < 0.01. 4.4. Heterogeneity analysis The previous sections discussed how AIIDEZ influences the environmental performance of the Chinese government and underlying mechanisms involved. However, existing research has not yet addressed the potential heterogeneity that may exist. This section explores the heterogeneous effects of AI on government environmental performance under different conditions. The model configuration for this analysis is as follows: $$\:{Y}_{i,t}=\alpha\:+{{\beta\:}_{1}\:Treat}_{i,t}\times\:Mod+{{\beta\:}_{2}\:Treat}_{i,t}+{\beta\:}_{3}Mod+{\gamma\:X}_{i,t}+{\mu\:}_{i}+{\lambda\:}_{t}+{\epsilon\:}_{i,t}$$ 4 In this model, \(\:Mod\) represents the moderating variables, which include Fiscal Autonomy 和 Highest Education of the Provincial Party Secretary . The other variables remain consistent with those in the baseline regression. 4.4.1. Fiscal autonomy Existing research indicates that fiscal autonomy plays a crucial role in environmental governance (He et al., 2024 ; Kostka & Nahm, 2017 ; Xu et al., 2024 ). Fiscal autonomy refers to the ratio of local government fiscal revenue to expenditure, reflecting the independence and flexibility of local governments in resource allocation and policy implementation. It determines their capacity to allocate resources during the implementation of AIIDEZ. The regression results are presented in Table 4 . The results in columns (1) and (2) show that the coefficients of the triple interaction term are significantly negative, indicating that fiscal autonomy negatively moderates the effect of AIIDEZ on reducing environmental pollution. Local governments with higher fiscal autonomy tend to focus more on economic development and innovation policies, possibly resulting in insufficient investment in environmental governance, which leads to a less significant reduction in environmental pollution compared to regions with lower fiscal autonomy. 4.4.2. Highest education of the provincial party secretary In China’s political system, the provincial party secretary, as the highest leader of local government, often plays a key role in formulating and implementing local policies. Research suggests that leaders’ educational background, particularly higher-level degrees, can influence their perceptions and support for technological innovation and environmental policies (Chen & Huang, 2021 ; Hong et al., 2021 ; Lu et al., 2020 ). Specifically, leaders with a higher level of education may have greater awareness and support for technological innovation and environmental governance, thereby affecting the local government’s policy implementation and effectiveness in achieving both AI and environmental protection goals. Conversely, leaders with a lower level of education may focus more on traditional economic growth objectives, neglecting environmental performance. Therefore, this study uses the highest educational level of the provincial party secretary as a moderating variable. Here, a diploma or below is coded as 1, a bachelor’s degree as 2, a master’s degree as 3, and a doctoral degree as 4. The aim is to explore the moderating effect of educational background on the relationship between AI innovation and environmental performance. The results in columns (3) to (4) show that the coefficient of the triple interaction term is significantly positive, indicating that the educational background of the provincial party secretary has a positive moderating effect on the reduction of environmental pollution through the establishment of AIIDEZ. Provincial party secretaries with a higher level of education are generally more sensitive and responsive to pilot reforms and capable of implementing more effective environmental protection policies while promoting technological innovation. This suggests that a higher level of education enhances the highest government leader’s understanding and ability to execute AIIDEZ policies, thereby intensifying the positive effects of AIIDEZ on environmental pollution governance. Table 4 Heterogeneity analysis M: Fiscal Autonomy M: Highest Education of the Provincial Party Secretary M ×Treat -0.187*** -0.127* 0.015** 0.023* (0.064) (0.069) (0.007) (0.013) Treat 0.020 -0.009 -0.122** -0.157*** (0.043) (0.041) (0.059) (0.045) M 0.027 0.210 0.007 0.003 (0.185) (0.210) (0.008) (0.006) Economic controls N Y N Y Meteorology controls N Y N Y Environmental pollution controls N Y N Y Year FE Y Y Y Y Province FE Y Y Y Y Within R² 0.127 0.314 0.108 0.311 Observation 360 360 360 360 Notes : Standard errors in parentheses are clustered at province level, * p < 0.10, ** p < 0.05, *** p < 0.01. 5. Concluding remarks This study delves into the impact of the establishment of China’s AIIDEZ on government environmental performance, focusing on the key role of government in environmental governance. Through an empirical analysis, we find that the establishment of AIIDEZ significantly enhances government environmental performance by promoting green innovation and strengthening environmental regulations. However, when local governments have higher fiscal autonomy, they are more likely to prioritize economic growth, which weakens their enthusiasm for environmental governance. In addition, government leaders with higher levels of education are more sensitive and proactive in responding to policy reforms, which allows them to drive environmental governance more effectively and improve environmental performance. This study makes several contributions to the existing literature. While prior research on environmental governance has largely focused on corporate environmental responsibility (Ambec & Lanoie, 2008 ; Porter & Linde, 1995 ) and the role of technological innovation in firm-level environmental performance (Ghisetti & Rennings, 2014 ; Horbach et al., 2012 ), our study shifts the analytical lens to the government level. We align with and extend the literature on technology-driven public governance (Margetts & Dunleavy, 2013 ; Meijer & Bolívar, 2016 ) by demonstrating how AI can enhance regulatory effectiveness and governance capabilities, ultimately improving environmental outcomes. Additionally, the findings contribute to the broader discourse on principal-agent problems in environmental policy implementation. Previous studies have highlighted the challenges local governments face in aligning national environmental goals with local economic priorities (Kostka & Hobbs, 2012 ; Rhodes, 2018 ). Our study provides empirical evidence that fiscal autonomy exacerbates these tensions, limiting the effectiveness of AI-driven environmental governance. This insight adds depth to existing research on decentralized governance and environmental policy execution (Berardo & Lubell, 2019 ; Oates, 2002 ), suggesting that AI adoption alone is insufficient without appropriate institutional constraints and incentives. By framing AI as a catalyst for governance capacity enhancement, this study refines theoretical understandings of digital government and environmental performance. While digital transformation has been widely explored in public administration (Dunleavy, 2006 ; Lips, 2019 ), its implications for environmental governance, particularly in emerging economies, remain underexamined. Our findings demonstrate that AI functions as both a policy instrument and an enforcement mechanism, bridging the gap between technological advancements and sustainable governance. From a policy perspective, this study highlights the necessity of complementary structural reforms to maximize AI’s governance potential. Given that fiscal autonomy may dilute environmental performance incentives, policymakers should consider environmental performance-linked fiscal transfers and AI-driven compliance monitoring to mitigate the trade-offs between economic growth and ecological sustainability. Furthermore, the observed correlation between leaders’ education levels and effective AI-driven environmental governance underscores the need for targeted investments in administrative capacity-building and leadership training. Future research should systematically examine the heterogeneous impacts of AIIDEZ across varying economic and policy contexts, identifying key determinants that shape AI’s efficacy in environmental governance. Beyond AI, the exploration of emerging technologies such as blockchain, big data, and the Internet of Things (IoT) could expand the theoretical and empirical understanding of technology-enabled governance frameworks. Moreover, the institutional mechanisms governing AI’s integration into environmental policy demand rigorous scrutiny, particularly concerning transparency, accountability, and ethical considerations in algorithmic decision-making. Addressing these dimensions will not only refine governance models but also enhance the strategic deployment of AI and related technologies in advancing sustainable environmental policies. In conclusion, This study offers a novel perspective on AI’s application in governmental environmental performance, shifting the discourse from corporate environmental responsibility to state-led ecological governance. By contributing to the literature on environmental governance, digital public administration, and policy implementation, it advances theoretical and empirical understandings of technology-driven governance. As emerging economies grapple with the dual imperatives of technological innovation and environmental sustainability, our findings offer vital insights for policymakers and scholars alike. This study lays a robust foundation for future research at the intersection of AI, public governance, and environmental policy, underscoring the imperative for an interdisciplinary approach to sustainable development. Declarations Disclosure statement of conflicts of interest The authors declare that they have no conflict of interest. Funding No funding was received for conducting this study. Author Contribution Z.L. designed the study, collected and analyzed the data, and wrote the initial draft of the manuscript. H.L. supervised the study, provided critical feedback, and revised the manuscript. Both authors reviewed and approved the final version of the manuscript. Data Availability The data that support the findings of this study are available from the corresponding author upon reasonable request. References Acemoglu D, Restrepo P (2018) Artificial intelligence, automation, and work. The economics of artificial intelligence: An agenda. University of Chicago Press, pp 197–236 Ambec S, Lanoie P (2008) Does it pay to be green? A systematic overview. Acad Manage Perspect, 45–62 Berardo R, Lubell M (2019) The ecology of games as a theory of polycentricity: Recent advances and future challenges. Policy Stud J 47(1):6–26 Cao G, Weng X, Xu M, Zhou L-A (2023) Hybrid Contracts, Multitasking, and Incentives: Theory and Evidence from China's Air Pollution Controls. Multitasking, and Incentives: Theory and Evidence from China's Air Pollution Controls (May 25, 2023) Chen J, Fleisher BM (1996) Regional income inequality and economic growth in China. J Comp Econ 22(2):141–164 Chen J, Huang C (2021) Policy Reinvention and Diffusion: Evidence from Chinese Provincial Governments. J Chin Polit Sci 26(4):723–741 Dell M (2010) The Persistent Effects of Peru's Mining Mita. Econometrica 78(6):1863–1903 Dong X, Zhou N, Zhao X, Yang S (2025) The impact of artificial intelligence on corporate green innovation: Can increasing quantity and improving quality go hand in hand? J Environ Manage 376:124439 Dunleavy P (2006) Digital era governance: IT corporations, the state, and e-government. Oxford University Press Ghisetti C, Rennings K (2014) Environmental innovations and profitability: How does it pay to be green? An empirical analysis on the German innovation survey. J Clean Prod 75:106–117 Ghobakhloo M, Fathi M (2021) Industry 4.0 and opportunities for energy sustainability. J Clean Prod 295:126427 Greenstone M, Hanna R (2014) Environmental regulations, air and water pollution, and infant mortality in India. Am Econ Rev 104(10):3038–3072 Greenstone M, He G, Jia R, Liu T (2022) Can technology solve the principal-agent problem? Evidence from China’s war on air pollution. Am Economic Review: Insights 4(1):54–70 Greenstone M, He G, Li S, Zou EY (2021) China’s war on pollution: Evidence from the first 5 years. Rev Environ Econ Policy 15(2):281–299 Guo Q, Peng Y, Luo K (2025) The impact of artificial intelligence on energy environmental performance: Empirical evidence from cities in China. Energy Econ 141:108136 He Q, Wei F, Zhang K, Zhong R, Kong F, Qi Y (2024) Fiscal decentralization, leader localization, and reduction of pollution and carbon emissions—empirical evidence from China's fiscal province-managing-county reform. J Environ Manage 360:121175 Hong T, Yu NN, Mao ZG, Zhang SH (2021) Government-driven urbanisation and its impact on regional economic growth in China. Cities, 117 , Article 103299. Horbach J, Rammer C, Rennings K (2012) Determinants of eco-innovations by type of environmental impact—The role of regulatory push/pull, technology push and market pull. Ecol Econ 78:112–122 Hu S, Wang AL, Du KR (2023) Environmental tax reform and greenwashing: Evidence from Chinese listed companies. Energy Econ 124:106873 Hussain M, Yang S, Maqsood US, Zahid RA (2024) Tapping into the green potential: The power of artificial intelligence adoption in corporate green innovation drive. Bus Strategy Environ 33(5):4375–4396 Jacobson LS, Lalonde RJ, Sullivan DG (1993) EARNINGS LOSSES OF DISPLACED WORKERS. Am Econ Rev 83(4):685–709 Jia R, Nie H (2017) Decentralization, collusion, and coal mine deaths. Rev Econ Stat 99(1):105–118 Jiang X, Li G, Fu W (2021) Government environmental governance, structural adjustment and air quality: A quasi-natural experiment based on the Three-year Action Plan to Win the Blue Sky Defense War. J Environ Manage 277:111470 Jin CF, Monfort A, Chen F, Xia N, Wu B (2024) Institutional investor ESG activism and corporate green innovation against climate change: Exploring differences between digital and non-digital firms. Technol Forecast Soc Chang, 200 , Article 123129. Kanabkaew T, Mekbungwan P, Raksakietisak S, Kanchanasut K (2019) Detection of PM2. 5 plume movement from IoT ground level monitoring data. Environ Pollut 252:543–552 Kluczek A, Zeglen P, Matusíková D (2021) The Use of Prospect Theory for Energy Sustainable Industry 4.0. Energies, 14 (22), Article 7694. Kostka G, Hobbs W (2012) Local energy efficiency policy implementation in China: bridging the gap between national priorities and local interests. China Q 211:765–785 Kostka G, Nahm J (2017) Central–local relations: Recentralization and environmental governance in China. China Q 231:567–582 Kou P, Han Y, Qi XY (2022) The operational mechanism and effectiveness of China's central environmental protection inspection: Evidence from air pollution. Socio-Economic Plann Sci 81:101215 Li HY, Lu ZQ, Zhang ZP, Tanasescu C (2025) How does artificial intelligence affect manufacturing firms' energy intensity? Energy Econ, 141 , Article 108109. Li K, Lin B (2017) Economic growth model, structural transformation, and green productivity in China. Appl Energy 187:489–500 Liang JQ, Langbein L (2015) Performance Management, High-Powered Incentives, and Environmental Policies in China. Int Public Manage J 18(3):346–385 Lips M (2019) Digital government: Managing public sector reform in the digital era. Routledge Liu WJ, Liu CX, Luo J, Liu F (2024) How does digital transformation promote total factor productivity? Strategy, technology, and application. Manag Decis Econ 45(5):2739–2750 Liu Y, Wang B, Song Z (2025) Promoting or inhibiting: The impact of artificial intelligence application on corporate environmental performance. Int Rev Financial Anal 97:103872 Liu ZF, Liu MD, Xia F, Zhang B (2024) Beyond technology: reducing environmental information asymmetry by centralizing environmental monitoring. Int Public Manage J 27(6):853–873 Lu J, Li B, Li H, Zhang XD (2020) Characteristics, exchange experience, and environmental efficiency of mayors: Evidence from 273 prefecture-level cities in China. J Environ Manage 255:109916 Luo Z, Li H (2018) The impact of Atmosphere Ten Articles policy on air quality in China. China Ind Econ 9:136–154 Margetts H, Dunleavy P (2013) The second wave of digital-era governance: a quasi-paradigm for government on the Web. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , 371 (1987), 20120382 Meijer A, Bolívar MPR (2016) Governing the smart city: a review of the literature on smart urban governance. Int Rev Admin Sci 82(2):392–408 Oates WE (2002) A reconsideration of environmental federalism. Recent advances in environmental economics. Edward Elgar Publishing Porter ME, Linde Cvd (1995) Toward a new conception of the environment-competitiveness relationship. J Economic Perspect 9(4):97–118 Razzaq A, Sharif A, Ozturk I, Yang XD (2023) Central inspections of environmental protection and transition for low-carbon Chinese cities: Policy intervention and mechanism analysis. Energy Econ 124:106859 Rhodes RAW (2018) Control and power in central-local government relations. Routledge Sun YP, Guan WM, Cao YN, Bao Q (2022) Role of green finance policy in renewable energy deployment for carbon Evidence from China. Renewable Energy 197:643–653 Tan YT, Mao XQ (2021) Assessment of the policy effectiveness of Central Inspections of Environmental Protection on improving air quality in China. J Clean Prod 288:125100 Wang A (2022) Chinese Environmental Law. China Q 249:282–284 Wang A, Luo K, Nie Y (2024) Can artificial intelligence improve enterprise environmental performance: Evidence from China. J Environ Manage 370:123079 Wang LK, Ma PP, Song Y, Zhang M (2023) How does environmental tax affect enterprises' total factor productivity? Evidence from the reform of environmental fee-to-tax in China. J Clean Prod, 413 , Article 137441. Wang Z, Zhang T, Ren X, Shi Y (2024) AI adoption rate and corporate green innovation efficiency: evidence from Chinese energy companies. Energy Econ 132:107499 Wei J, Li ZQ, Cribb M, Huang W, Xue WH, Sun L, Guo JP, Peng YR, Li J, Lyapustin A, Liu L, Wu H, Song YM (2020) Improved 1 km resolution PM2.5 estimates across China using enhanced space-time extremely randomized trees. Atmos Chem Phys 20(6):3273–3289 Wei J, Li ZQ, Lyapustin A, Sun L, Peng YR, Xue WH, Su TN, Cribb M (2021) Reconstructing 1-km-resolution high-quality PM2.5 data records from 2000 to 2018 in China: spatiotemporal variations and policy implications. Remote Sens Environ 252:112136 Wu RX, Hu P (2019) Does the Miracle Drug of Environmental Governance Really Improve Air Quality? Evidence from China's System of Central Environmental Protection Inspections. Int J Environ Res Public Health, 16 (5), Article 850. Xu H, Li Y, Lin W, Li Y (2024) Government fiscal decentralization and haze and carbon reduction: Evidence from the fiscal Province-Managing-County reform. Environ Res 252:119020 Yang C-H (2022) How artificial intelligence technology affects productivity and employment: firm-level evidence from Taiwan. Res Policy 51(6):104536 Yang L, Lin Y, Wang J, Peng F (2024) Achieving Air Pollution Control Targets with Technology–Aided Monitoring: Better Enforcement or Localized Efforts? Am Economic Journal: Economic Policy 16(4):280–315 Yang M, Chou S-Y (2018) The impact of environmental regulation on fetal health: evidence from the shutdown of a coal-fired power plant located upwind of New Jersey. J Environ Econ Manag 90:269–293 Yao J, Zhang K, Guo L, Feng X (2024) How does artificial intelligence improve firm productivity? Based on the perspective of labor skill structure adjustment. J Manage World 40(2):101–116 Yi F, Niu ZH, Zhong SB (2025) Does information and communication technology reduce carbon emissions in China? Evidence from the quasi-natural experiment of the Broadband China pilot policy. J Environ Planning Manage 68(1):1–27 Ying Y, Cui XY, Jin SY (2023) Artificial Intelligence and Green Total Factor Productivity: The Moderating Effect of Slack Resources. Systems, 11 (7), Article 356. Zhang DY (2022) Green financial system regulation shock and greenwashing behaviors: Evidence from Chinese firms. Energy Econ 111:106064 Zhang GL (2023) Regulatory-driven corporate greenwashing: Evidence from ?low-carbon city? pilot policy in China. Pac-Basin Financ J, 78 , Article 101951. Zhang K, Kou ZX, Zhu PH, Qian XY, Yang YZ (2025) How does AI affect urban carbon emissions? Quasi-experimental evidence from China's AI innovation and development pilot zones. Econ Anal Policy 85:426–447 Zhang P, Wu JN (2018) Impact of mandatory targets on PM2.5 concentration control in Chinese cities. J Clean Prod 197:323–331 Zhou W, Zhuang Y, Chen Y (2024) How does artificial intelligence affect pollutant emissions by improving energy efficiency and developing green technology. Energy Econ 131:107355 Zhu XF, Wang Y (2024) Credible signaling to promote local compliance: Evidence from China's multiwave inspection of environmental protection. Public Adm Zhuang MX (2024) Supervising Local Cadres in China: The Quest for Authoritarian Accountability. Politics Soc 52(3):452–485 Additional Declarations No competing interests reported. Supplementary Files Appendix.docx 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-7322815","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":552284809,"identity":"2e7c6876-bb4f-4fa3-adf1-9a8f700cccc0","order_by":0,"name":"Zhiwei Liu","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Zhiwei","middleName":"","lastName":"Liu","suffix":""},{"id":552284810,"identity":"1971ea71-21fe-4202-a2e6-8ceb79dcff38","order_by":1,"name":"Haiming Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzklEQVRIiWNgGAWjYBACefb+7z8SKv7LsbE3EKnFsOeAgcSHM8zG/DwHiLXmRoKB5MwW5kTJGQlE6mCckZBgzNvAlmBw8/HGGww1NtEEtbDzPDiQzLuDJ8/gdlqxBcOxtNwGgra0JzYc5j0jUWxwO8dMgrHhMGEtDAeSGZt52wwSN9w8Q6yWE2nMjDPbEhJnzuAhUothzxk2hg9nDgADGeiXBGL8Is/ew8aQUHEAGJWHN974UGNDhMOQgIFEAinKIVpI1TEKRsEoGAUjAwAA739Cf6VO77sAAAAASUVORK5CYII=","orcid":"","institution":"Wenzhou Polytechnic","correspondingAuthor":true,"prefix":"","firstName":"Haiming","middleName":"","lastName":"Liu","suffix":""}],"badges":[],"createdAt":"2025-08-08 02:53:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7322815/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7322815/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97249722,"identity":"247bf1a3-9b5c-48d9-9e3a-1f0d8718b0c8","added_by":"auto","created_at":"2025-12-02 13:13:17","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":247124,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/5e538560eb879f354c80d2d0.docx"},{"id":97160378,"identity":"1a84eb3f-54a2-4de6-a037-3e3e0fd62dac","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3978,"visible":true,"origin":"","legend":"","description":"","filename":"bba9a5bf5d4c4eaeacbb471b78cf9493.json","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/83f474e8658a1815b37bee8b.json"},{"id":97248840,"identity":"b717a745-b7d8-436e-8185-8c43a767554b","added_by":"auto","created_at":"2025-12-02 13:07:34","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77447,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/bb90ed0b52d6382e50c6c50c.docx"},{"id":97248854,"identity":"9ae0493c-8db3-4974-8fb7-ccf98140f55d","added_by":"auto","created_at":"2025-12-02 13:07:40","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":150647,"visible":true,"origin":"","legend":"","description":"","filename":"bba9a5bf5d4c4eaeacbb471b78cf94931enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/d6d7b651720cdaee2e224d75.xml"},{"id":97160384,"identity":"329a8b76-fbc0-43e1-91f3-e166ba022f20","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39078,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/6bcd9341fab1963e1eafc599.png"},{"id":97160382,"identity":"cdb7ac27-761a-45fe-822e-8c787d370403","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12488,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/c6c68e61aed38231c484d61b.png"},{"id":97160385,"identity":"c467e517-5053-44df-8d9c-e7957e93a001","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":149181,"visible":true,"origin":"","legend":"","description":"","filename":"bba9a5bf5d4c4eaeacbb471b78cf94931structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/6386bbef49f0393d6f64e99a.xml"},{"id":97160387,"identity":"a8311870-d943-4d19-b90a-f0217f736288","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"html","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":155342,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/5c06e64b845391f819e731e1.html"},{"id":97160377,"identity":"d3de7c10-647b-415d-b70a-9df38f5395a2","added_by":"auto","created_at":"2025-12-01 12:27:29","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":96515,"visible":true,"origin":"","legend":"\u003cp\u003eMap of AIIDEZ Progress\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eNotes:\u003c/em\u003e In cases where two or more cities within the same province are approved as pilot cities, resulting in ambiguity regarding the starting time for assigning the Time variable to the province, we use the time of approval for the first city approved within that province as the baseline for assigning the Time variable to the province.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/808917cb412d85ab6fb69100.jpeg"},{"id":97160379,"identity":"6c7c4e8a-72f7-4abf-9841-cd1412599667","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":43735,"visible":true,"origin":"","legend":"\u003cp\u003eResults of the parallel trend test\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/5f3b605eb53008151a0a718b.png"},{"id":109405092,"identity":"8dd189e1-bfcb-434e-bc81-a63fe7b2d896","added_by":"auto","created_at":"2026-05-17 12:55:00","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":510839,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/f4bb4ff4-add7-4e8d-9491-7c6fef6c377d.pdf"},{"id":97160381,"identity":"b658e0f0-beac-4fea-8b7b-786744bac33e","added_by":"auto","created_at":"2025-12-01 12:27:30","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":77447,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-7322815/v1/8a174f04211bcb200961d9f4.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Technological Transformation of Environmental Governance: Evidence from a Quasi-Natural Experiment in China","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn recent years, environmental issues have emerged as a central focus of global concern. In this regard, China is at the forefront due to its severe pollution, which poses a substantial challenge to the sustainable development of both its economy and society. While China has achieved remarkable economic growth over the past few decades, the accompanying environmental pressures have intensified (Chen \u0026amp; Fleisher, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e1996\u003c/span\u003e; Li \u0026amp; Lin, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In 2010, more than two-thirds of the world\u0026rsquo;s 50 most polluted cities were in China, and the country\u0026rsquo;s pollution emissions ranked among the highest globally. Despite the Chinese government\u0026rsquo;s proactive policy initiatives to address environmental degradation, local governments continue to face significant challenges in policy implementation, particularly in relation to conflicts of interest and information asymmetry, which substantially undermine the effectiveness of environmental governance (Cao et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Greenstone et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In this context, emerging technologies, particularly artificial intelligence (AI), are increasingly viewed as promising tools for alleviating information asymmetries and enhancing the effectiveness of environmental governance. However, whether these technologies can effectively address the principal-agent problems inherent in environmental governance remains an unresolved and complex issue, warranting further scholarly exploration.\u003c/p\u003e\u003cp\u003eAI, a key driver of technological and industrial transformation, has the potential to significantly enhance productivity, foster innovation, and stimulate economic growth (Acemoglu \u0026amp; Restrepo, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Li et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; W. J. Liu et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Yao et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Existing research has predominantly focused on the impact of AI on corporate environmental performance, with numerous studies demonstrating the crucial role of AI technologies in improving the environmental outcomes of firms (Ying et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Zhou et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). On one hand, AI directly contributes to environmental performance by optimizing production processes, improving resource efficiency, and promoting green innovation. AI not only enhances the quantity of corporate green innovations, but also increases their innovation efficiency (Dong et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Hussain et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Z. Wang et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Intelligent manufacturing and energy distribution systems allow AI to significantly reduce energy waste, thereby optimizing environmental performance (Ghobakhloo \u0026amp; Fathi, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Kluczek et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Furthermore, AI offers distinct advantages in intelligent scheduling and energy demand forecasting, enabling firms to effectively reduce carbon dioxide emissions and providing a technological foundation for achieving green development (Kanabkaew et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Sun et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, this governance model is not without controversy. External regulatory pressures may encourage firms to engage in symbolic environmental actions rather than substantive improvements. While green finance policies have accelerated corporate green transformation, they often come with financial constraints, prompting some high-polluting firms to resort to \u0026ldquo;greenwashing.\u0026rdquo; These firms may superficially comply with environmental standards without implementing genuine green innovations (Hu et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Yi et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Zhang, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zhang, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGovernment attention and involvement (as reflected in governmental documents and policies) represent the most direct factors influencing environmental quality (Zhang \u0026amp; Wu, \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). In recent years, both domestic and international scholars have increasingly focused on evaluating the effectiveness of government environmental governance. Research indicates that air pollution regulation in developing countries can effectively reduce particulate matter in the air, thereby accelerating the process of improving environmental quality (Greenstone \u0026amp; Hanna, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Luo \u0026amp; Li, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Yang \u0026amp; Chou, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Within the realm of environmental governance, the application of technologies, particularly automated monitoring systems, is crucial for enhancing governance effectiveness and ensuring compliance. For instance, the introduction of automated air pollution monitoring has not only effectively curbed local government manipulation of data, but also significantly improved the accuracy of air quality reports (Greenstone et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Greenstone et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The use of such technologies has brought real-time data and transparency, thereby advancing policy implementation and reducing information biases and deficiencies in the governance process. Studies have shown that automated monitoring systems have prompted local governments to respond swiftly to pollution hotspots, substantially reducing pollution levels in monitored areas (Yang et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Moreover, as governments continue to strengthen pollution control policies, environmentally supportive policies backed by technology have effectively reduced pollutant emissions and improved air quality (Cao et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Jiang et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, despite the pivotal role of technology in enhancing environmental governance, issues such as technological lag and data manipulation persist, necessitating the safeguarding of the fairness and effectiveness of technology application within institutional frameworks (Z. F. Liu et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Consequently, while AI serves as a key driver of technological and societal transformation, its impact on government environmental performance remains uncertain. Furthermore, insufficient attention is given to the actual occurrence and mechanisms of this technology, which makes it difficult for both scholars and policymakers to accurately assess the real effectiveness of AI in improving government environmental performance. It also complicates comprehensively evaluating the potential implications of such reforms on central-local government relations.\u003c/p\u003e\u003cp\u003eConsidering this, this study focuses on the pilot cities of the three batches of \u0026ldquo;Artificial Intelligence Innovation Development Experimental Zones\u0026rdquo;(AIIDEZ) established in China from 2019 to 2021. Using environmental performance as the key outcome variable, this research systematically examines the strategic choices made by local governments in different contexts following the establishment of these AI innovation zones. Unlike previous studies that focused on corporate-level environmental performance, this research centers on the role of government in driving environmental improvements, positing governmental policy guidance and governance capacity as the critical drivers of environmental enhancement. By constructing a panel dataset of 30 Chinese provinces from 2011 to 2022 and applying multiple robustness checks, we rigorously analyze the impact of AIIDEZ on environmental performance and explore the underlying mechanisms. The empirical results demonstrate that this reform pilot has significantly improved government environmental performance by promoting green innovation and strengthening environmental regulation. Furthermore, the study reveals that government leaders with a higher level of education are more sensitive to and proactive in responding to the pilot reforms. However, when local governments have relatively greater fiscal autonomy, officials tend to prioritize economic development, thereby diminishing the actual effectiveness of the reform pilot.\u003c/p\u003e\u003cp\u003eExisting research on government environmental governance and air quality lays the foundation for our study and opens avenues for further exploration. The innovative contributions of our research are as follows: First, it deepens understanding of the consequences of AI application. Previous studies focused on the impact of AI applications on the economy (Yang, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Yao et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and corporate environmental performance (Guo et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; A. Wang et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), often overlooking the crucial role of government as a key factor in the application of AI. By examining the relationship between AI application and government environmental performance, this study expands understanding of AI\u0026rsquo;s broader implications. Second, this research contributes to the study of environmental performance in emerging economies. While existing literature primarily explores the impact of internal technological advancements on the environmental performance of emerging economies (Jin et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), this study offers a novel perspective by investigating the effect of AI application on government environmental performance through the lens of internal principal-agent incentives, enriching knowledge on environmental performance in emerging economies. Finally, from a qualitative perspective, this study analyzes the \u0026ldquo;Chinese-style\u0026rdquo; logic and mechanisms underlying the significant outcomes of AIIDEZ, providing new insights for governments to continue implementing environmental governance policies and achieving harmonious coexistence between humans and nature.\u003c/p\u003e"},{"header":"2. AIIDEZ in China","content":"\u003cp\u003eDriven by globalization and modernization, government organizations often face multiple competing tasks in A real-world governance context. For instance, when advancing regional economic development with a focus on high-quality growth, governments frequently encounter the dilemma of balancing development and protection. Economic growth typically relies on resource-intensive and high-pollution industries, the expansion of which inevitably leads to environmental pollution, ecological degradation, and a decline in social welfare. Improving environmental performance often necessitates costly technological upgrades or industrial restructuring, placing local governments under pressure to meet short-term economic goals while simultaneously addressing long-term environmental sustainability challenges.\u003c/p\u003e\u003cp\u003eIn 2003, Chinese President Hu Jintao introduced the \u0026ldquo;Scientific Outlook on Development,\u0026rdquo; emphasizing the need to balance economic growth with environmental protection. Since then, promoting sustainable development has been a central component of China\u0026rsquo;s national strategy. In practice, environmental governance in China follows a primary \u0026ldquo;command and control\u0026rdquo; model, which is similar to the target-based system under the planned economy. At the beginning of each year, the central government sets nationwide targets for the reduction of various pollutants. These targets are then allocated to provinces, which then assign them to municipalities and counties. To incentivize compliance, the achievement of environmental governance goals is directly linked to the political advancement of key government leaders (Liang \u0026amp; Langbein, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In 2007, the State Council issued \u0026ldquo;Measures for the Assessment of Total Pollutant Emissions,\u0026rdquo; implementing a \u0026ldquo;one strike and you\u0026rsquo;re out\u0026rdquo; rule for government leaders who failed to meet their pollution reduction targets. Such leaders were ineligible for top performance evaluations. Furthermore, the central government revised the \u0026ldquo;Environmental Protection Law of the People's Republic of China\u0026rdquo; in 2014, officially establishing the legal responsibility of governments at all levels for environmental quality within their jurisdictions, and clearly defining emission standards for various pollutants (Wang, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eDespite the central government\u0026rsquo;s long-standing emphasis on environmental governance, in practice, government negligence in environmental regulation is often difficult to detect and correct in a timely manner. To address these issues, the central government has gradually introduced a series of environmental supervision systems, including environmental protection inspections characterized by \u0026ldquo;enterprise supervision\u0026rdquo; (referred to as \u0026ldquo;environmental inspections\u0026rdquo;) and central environmental protection inspections marked by \u0026ldquo;government supervision\u0026rdquo; (referred to as \u0026ldquo;central environmental inspections\u0026rdquo;) (Zhu \u0026amp; Wang, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zhuang, \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), However, while these reforms, which focus on \u0026ldquo;enterprise supervision\u0026rdquo; and \u0026ldquo;government supervision,\u0026rdquo; have succeeded in partially centralizing environmental power, they have not fundamentally altered the issue of the excessive centralization of environmental authority. Some studies have found that although environmental performance improves significantly during inspection periods, the policy effects immediately diminish once the inspections are over, and the government\u0026rsquo;s enforcement bias persists (Kou et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Razzaq et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Tan \u0026amp; Mao, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Wu \u0026amp; Hu, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)。\u003c/p\u003e\u003cp\u003eIn the face of this complex situation, AI, a disruptive innovation, offers a potential pathway for enhancing environmental governance capabilities. By integrating technologies such as big data analytics, machine learning, and the Internet of Things (IoT), AI can provide efficient technical support across various domains including environmental monitoring, resource optimization, and pollution control. Through AI, local governments can achieve more precise, dynamic, and intelligent environmental governance. Moreover, as a new production factor, AI can improve management and organizational efficiency, facilitating the intelligent and green transformation of economic activities. This enables the transition of China\u0026rsquo;s traditional industries towards higher-end, smarter, and greener models, thereby alleviating the conflict between economic development and environmental protection. Against this backdrop, in 2019, the Ministry of Science and Technology of China issued the \u0026ldquo;Guidelines for the Construction of National New Generation AI Innovation Development Experimental Zones.\u0026rdquo; To date, 18 cities including Beijing have been designated as AIIDEZ, The list is provided in Appendix A. The establishment of these zones aims to promote the application and industrialization of AI technologies through policy support and resource allocation, serving as a key driver for high-quality economic development. What are the policy effects of these zones? Have they significantly improved government environmental performance? What are the underlying mechanisms? Clarifying these questions is of crucial practical significance for innovating AI development models, defining AI\u0026rsquo;s future trajectory, enhancing environmental performance, and driving high-quality economic development.\u003c/p\u003e"},{"header":"3. Empirical design","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e3.1. Dependent variable\u003c/h2\u003e\u003cp\u003eThe primary dependent variable in this study is government environmental performance. To measure this, we use the \u003cem\u003ePM2.5\u003c/em\u003e data from the China High Air Pollutants (CHAP) dataset, which is published by the National Qinghai-Tibet Plateau Science Data Center (Wei et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wei et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In the robustness checks section (section \u003cspan refid=\"Sec10\" class=\"InternalRef\"\u003e4.2\u003c/span\u003e), we replace the dependent variable with global surface \u003cem\u003ePM2.5\u003c/em\u003e concentration data (\u0026micro;g/m\u0026sup3;) from the University of Washington St. Louis campus and \u003cem\u003eAQI\u003c/em\u003e data from the National Urban Air Quality Real-time Release Platform of the China National Environmental Monitoring Center. This substitution ensures the robustness of our study.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e3.2. Independent variables\u003c/h2\u003e\u003cp\u003eThe primary explanatory variable in this study is a binary policy indicator variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Treat}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e。If province \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003e is approved as an AIIDEZ in year \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e, the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Treat}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e for the corresponding province is assigned a value of 1 for year \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e and subsequent years; otherwise, it is assigned a value of 0. The Ministry of Science and Technology of China established pilot cities in three batches in 2019, 2020, and 2021. Since some pilot cities were approved toward the end of the year, we account for the lag effect of the policy by assigning cities approved after September to the following year. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the timeline of the AIIDEZ across different regions in China. As Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows, reform implementation exhibits notable spatiotemporal variation. The program launched in Beijing and Shanghai in 2019, gradually expanding to other eastern provinces and the central-western regions thereafter. The deep red areas represent regions that implemented AIIDEZ earlier, while the lighter red areas indicate those that adopted the reform later. The gray-shaded areas reflect the limitations of the reform\u0026rsquo;s coverage. This spatial distribution of AIIDEZ highlights the dynamic progression of AI development in China.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e3.3. Control variables\u003c/h2\u003e\u003cp\u003e\u003cstrong\u003eUrban socioeconomic data\u003c/strong\u003e\u003cp\u003eIn addition, we control for various urban socioeconomic factors including per capita GDP, public budget revenue, and resident population. The data primarily comes from the China Statistical Yearbook.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eUrban environmental pollution data\u003c/strong\u003e\u003cp\u003eThe environmental performance of a province may be influenced by the local level of environmental pollution. Therefore, we control for total wastewater discharge, sulfur dioxide emissions from exhaust gases, the generation of general industrial solid waste, and total electricity consumption. The data mainly comes from the China Statistical Yearbook and the China Environmental Protection Statistical Yearbook.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eUrban meteorological data\u003c/strong\u003e\u003cp\u003eThe environmental performance of a province may also be affected by its geographical factors. Therefore, we control for various covariates related to a province\u0026rsquo;s climate characteristics including average temperature, annual cumulative precipitation, and average wind speed. These data are processed by the National Centers for Environmental Information (NCEI) under the National Oceanic and Atmospheric Administration (NOAA). Descriptive statistics for the key variables are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e3.4. Model specifications\u003c/h2\u003e\u003cp\u003eThis study treats the establishment of AIIDEZ in China as a quasi-natural experiment and rigorously evaluates the impact of these pilot zones on local environmental performance using a difference-in-differences (DID) model. Given that the implementation of AIIDEZ is staggered over time, it is not possible to create a single uniform dummy variable to reflect the establishment time across all regions. Therefore, we classify provinces with established AI innovation zones as the treatment group, and those without as the control group. In addition, we identify the specific start time for each province\u0026rsquo;s participation in the pilot program to define the policy dummy variable. The empirical model structure reflecting this setup is as follows:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:{PM2.5}_{i,t}=\\alpha\\:+{\\beta\\:\\:Treat}_{i,t}+{\\gamma\\:X}_{i,t}+{\\mu\\:}_{i}+{\\lambda\\:}_{t}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eIn this model, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003e denotes the province and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e the year. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{PM2.5}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e represents environmental performance. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Treat}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e is the primary explanatory variable, which indicates whether province \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003e was designated as an AIIDEZ in year \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e. This study focuses on the coefficient \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Treat}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e, as it reflects the net effect of establishing an AIIDEZ on local environmental performance. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{X}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e includes a range of province-specific factors, covering natural climate variables and socioeconomic factors. In the empirical framework, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\mu\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003e represents province fixed effects, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\lambda\\:}_{t}\\)\u003c/span\u003e\u003c/span\u003e year fixed effects, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\epsilon\\:}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e the random disturbance term.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDescriptive statistics of major variables.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMean\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSD\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMin\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eMax\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog PM2.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e3.641\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.374\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2.664\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e4.535\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTreat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.117\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.321\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog PerGDP\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e10.87\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.461\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e9.682\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e12.15\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Fiscal revenue\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e7.650\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.844\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.023\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e9.554\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Population\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e8.208\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.741\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6.342\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e9.448\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e2.547\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.499\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.890\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3.255\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Precipitation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e6.849\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.534\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4.993\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e7.843\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Wind\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e2.185\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.170\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.674\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e2.661\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Waste water\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e12.05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.819\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e9.966\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e13.75\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Solid waste\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e2.984\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.368\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-2.207\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e5.208\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog SO2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e8.986\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.062\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.142\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e10.86\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLog Electricity consumption\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e7.435\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.696\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5.222\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e8.971\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Empirical results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e4.1. Baseline results\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e provides a detailed description of the impact of the establishment of AIIDEZ on environmental performance. All Variables Used in Baseline Result listed in Appendix B. Model 1 estimates the simple binary relationship between the AI innovation development zones and environmental performance. Model 2 includes macroeconomic control variables at the province level, and Model 3 further incorporates meteorological control variables. Model 4, as the comprehensive model, additionally includes environmental pollution data at the province level. All models show that the coefficient of \u003cem\u003eTreat\u003c/em\u003e is significantly negative at the 1% significance level. Focusing on Model 4, the estimated values indicate that the establishment of experimental zones has led to an average reduction of 8% in \u003cem\u003ePM2.5\u003c/em\u003e levels, suggesting the effective role of AI innovation zones in reducing \u003cem\u003ePM2.5\u003c/em\u003e. While these findings provide strong evidence, the specific mechanisms through which AIIDEZ achieves these environmental outcomes remain uncertain. Further investigation of these potential pathways will deepen our understanding of the operational dynamics of this policy.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eBaseline results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e\u003cp\u003e\u003cem\u003eDV: PM2.5 (Log)\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eTreat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.088***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.095***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.094***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.080***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(0.025)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.026)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.024)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.016)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEconomic controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMeteorology controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEnvironmental pollution controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYear FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eProvince FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithin R\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.102\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.132\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.143\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.303\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eObservation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNotes\u003c/em\u003e: Standard errors in parentheses are clustered at province level, \u003csup\u003e*\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.10, \u003csup\u003e**\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u003csup\u003e***\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e4.2. Parallel trend and robustness tests\u003c/h2\u003e\u003cp\u003eThe validity of the DID estimation relies on the parallel trend assumption. This assumption posits that the provinces that implemented AIIDEZ exhibit no systematic differences in the \u003cem\u003ePM2.5\u003c/em\u003e concentration change trend compared to the provinces that did not implement the experimental zones before the introduction of the pilot program. To test this assumption, this study follows the event study method introduced by (Jacobson et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) to conduct a parallel trends test on the model. The model is set as follows:\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{i,t}=\\alpha\\:+{\\sum\\:_{k\\ge\\:-5}^{k}\\beta\\:\\:Treat}_{i,{t}_{0}+k}+{\\gamma\\:X}_{i,t}+{\\mu\\:}_{i}+{\\lambda\\:}_{t}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eIn this model, the variable \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Treat}_{i,{t}_{0}+k}\\)\u003c/span\u003e\u003c/span\u003e represents the event window dummy variable indicating the periods before and after the implementation of the AIIDEZ policy. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{0}\\)\u003c/span\u003e\u003c/span\u003e is the year the province completed the establishment of AIIDEZ, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{0}+k\\)\u003c/span\u003e\u003c/span\u003e refers to the various years before and after the policy pilot. The value of k takes on values of -5, -4, -3, -2, -1, 0, 1, 2, and 3, with the study introducing dummy variables for the five years prior to the pilot. Any year greater than or equal to five years is grouped into the five-year category. The study uses the year immediately preceding the establishment of AIIDEZ as the reference group, with other control variables and the baseline model remaining similar.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presents the parameter estimates and their corresponding confidence intervals. As the figure shows, for periods where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:k\u0026lt;0\\)\u003c/span\u003e\u003c/span\u003e, the value of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e does not significantly deviate from zero, indicating that prior to the reform, there were no significant differences in the trends of \u003cem\u003ePM2.5\u003c/em\u003e concentrations between provinces that implemented AIIDEZ and those that did not. This pattern satisfies the parallel trends assumption. In contrast, at \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:k=0\\)\u003c/span\u003e\u003c/span\u003e, the value of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e deviates notably from zero, showing a sharp decline. In subsequent periods, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e gradually decreases. This suggests that the implementation of AIIDEZ significantly slowed the increase in \u003cem\u003ePM2.5\u003c/em\u003e concentrations in the provinces that adopted the policy. Over time, this effect strengthens.\u003c/p\u003e\u003cp\u003eIn addition, we conducted a series of robustness checks to address concerns regarding measurement errors and potential omitted variables. These checks included analyzing different subsamples, considering alternative time division standards, accounting for the impact of other environmental policies during the policy period, and using alternative dependent variables. The results remained consistent with the baseline findings. Further details are provided in Appendix C.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e4.3. Mechanism analysis\u003c/h2\u003e\u003cp\u003eAs the government\u0026rsquo;s environmental governance functions evolve, it is necessary to further explore the underlying mechanisms through which PM2.5 concentrations are reduced. To this end, we delve into the potential channels through which AIIDEZ reduce PM2.5 concentrations, focusing on green innovation and the intensity of environmental regulations. Based on (Dell, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), the model is set as follows:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:Med=\\alpha\\:+{\\beta\\:\\:Treat}_{i,t}+{\\gamma\\:X}_{i,t}+{\\mu\\:}_{i}+{\\lambda\\:}_{t}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eWhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Med\\)\u003c/span\u003e\u003c/span\u003e represents the mechanism variables, which are proxied by \u003cem\u003eGreen Innovation\u003c/em\u003e and the \u003cem\u003eIntensity of Environmental Regulations\u003c/em\u003e. The other variables are set in accordance with the baseline model.\u003c/p\u003e\u003cdiv id=\"Sec12\" class=\"Section3\"\u003e\u003ch2\u003e4.3.1. Green innovation\u003c/h2\u003e\u003cp\u003eTheoretically, AIIDEZ can reduce \u003cem\u003ePM2.5\u003c/em\u003e concentrations by promoting green innovation. To explore this potential mechanism, this study uses the number of green patent applications and granted patents from the China National Research Data Service Platform (CNRDS) as proxies for green innovation. The regression analysis results in columns (1)\u0026ndash;(4) of Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e indicate that AIIDEZ effectively enhances local green innovation, consistent with theoretical expectations. This relationship may be attributed to the fact that AIIDEZ fosters green innovation through policy support, technological innovation, and industrial transformation, thereby reducing \u003cem\u003ePM2.5\u003c/em\u003e concentrations. Policy incentives have accelerated the growth of green patents, AI technology has improved the efficiency of green technologies, and industrial transformation has driven the green upgrading of high-polluting industries. These factors have worked synergistically to improve environmental quality. In summary, the study confirms the key role of green innovation in effectively reducing environmental pollution, further emphasizing the important mechanism through which AIIDEZs enhance air quality by promoting green innovation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\u003ch2\u003e4.3.2. Environmental regulation intensity\u003c/h2\u003e\u003cp\u003eExisting research has shown that a key factor contributing to environmental pollution is the government\u0026rsquo;s tendency to lower environmental regulations in pursuit of economic benefits, resulting in \u0026ldquo;local regulatory bias\u0026rdquo; (Jia \u0026amp; Nie, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). With the establishment of AIIDEZ, the government's environmental regulatory intensity and resource allocation in environmental governance may change. Given that the government\u0026rsquo;s environmental regulatory intensity is to some extent a reflection of the costs associated with pollution control, the stronger the environmental regulations, the higher the government\u0026rsquo;s pollution control costs. Therefore, we measure environmental regulation by the ratio of government industrial pollution control investment to industrial added value. The regression results in columns (5) and (6) of Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e show that AIIDEZ significantly enhances the government\u0026rsquo;s environmental regulatory intensity. This suggests that the government may optimize environmental governance by enhancing regulatory precision, improving decision-making efficiency, and reducing governance costs. With the help of AI technology, the government can achieve accurate pollution source identification and real-time monitoring, thereby improving the scientific and transparent nature of governance, and reducing information asymmetry and regulatory blind spots in traditional governance models. This shift alleviates the problem of local governments\u0026rsquo; \u0026ldquo;regulatory bias\u0026rdquo; and provides a new path for promoting more efficient and sustainable environmental governance.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eThe mechanism of AIIDPZ alleviates local PM2.5.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e\u003cp\u003e\u003cem\u003eDV: Green Innovation\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u003cp\u003e\u003cem\u003eDV: Environmental Regulation Intensity\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e(5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e(6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eTreat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e6482.485***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5252.981***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e6147.204***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5687.539***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.002***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.002***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(1734.179)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(1611.412)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(1726.790)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(1657.605)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e(0.001)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e(0.001)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEconomic controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMeteorology controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEnvironmental pollution controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYear FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eProvince FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithin R\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.120\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.326\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.178\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.325\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.042\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.122\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eObservation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eNotes\u003c/em\u003e: Standard errors in parentheses are clustered at province level, \u003csup\u003e*\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.10, \u003csup\u003e**\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u003csup\u003e***\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e4.4. Heterogeneity analysis\u003c/h2\u003e\u003cp\u003eThe previous sections discussed how AIIDEZ influences the environmental performance of the Chinese government and underlying mechanisms involved. However, existing research has not yet addressed the potential heterogeneity that may exist. This section explores the heterogeneous effects of AI on government environmental performance under different conditions. The model configuration for this analysis is as follows:\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{i,t}=\\alpha\\:+{{\\beta\\:}_{1}\\:Treat}_{i,t}\\times\\:Mod+{{\\beta\\:}_{2}\\:Treat}_{i,t}+{\\beta\\:}_{3}Mod+{\\gamma\\:X}_{i,t}+{\\mu\\:}_{i}+{\\lambda\\:}_{t}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eIn this model, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Mod\\)\u003c/span\u003e\u003c/span\u003e represents the moderating variables, which include \u003cem\u003eFiscal Autonomy\u003c/em\u003e和\u003cem\u003eHighest Education of the Provincial Party Secretary\u003c/em\u003e. The other variables remain consistent with those in the baseline regression.\u003c/p\u003e\u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\u003ch2\u003e4.4.1. Fiscal autonomy\u003c/h2\u003e\u003cp\u003eExisting research indicates that fiscal autonomy plays a crucial role in environmental governance (He et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Kostka \u0026amp; Nahm, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Xu et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Fiscal autonomy refers to the ratio of local government fiscal revenue to expenditure, reflecting the independence and flexibility of local governments in resource allocation and policy implementation. It determines their capacity to allocate resources during the implementation of AIIDEZ. The regression results are presented in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. The results in columns (1) and (2) show that the coefficients of the triple interaction term are significantly negative, indicating that fiscal autonomy negatively moderates the effect of AIIDEZ on reducing environmental pollution. Local governments with higher fiscal autonomy tend to focus more on economic development and innovation policies, possibly resulting in insufficient investment in environmental governance, which leads to a less significant reduction in environmental pollution compared to regions with lower fiscal autonomy.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003e4.4.2. Highest education of the provincial party secretary\u003c/h2\u003e\u003cp\u003eIn China\u0026rsquo;s political system, the provincial party secretary, as the highest leader of local government, often plays a key role in formulating and implementing local policies. Research suggests that leaders\u0026rsquo; educational background, particularly higher-level degrees, can influence their perceptions and support for technological innovation and environmental policies (Chen \u0026amp; Huang, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Hong et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Lu et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Specifically, leaders with a higher level of education may have greater awareness and support for technological innovation and environmental governance, thereby affecting the local government\u0026rsquo;s policy implementation and effectiveness in achieving both AI and environmental protection goals. Conversely, leaders with a lower level of education may focus more on traditional economic growth objectives, neglecting environmental performance. Therefore, this study uses the highest educational level of the provincial party secretary as a moderating variable. Here, a diploma or below is coded as 1, a bachelor\u0026rsquo;s degree as 2, a master\u0026rsquo;s degree as 3, and a doctoral degree as 4. The aim is to explore the moderating effect of educational background on the relationship between AI innovation and environmental performance. The results in columns (3) to (4) show that the coefficient of the triple interaction term is significantly positive, indicating that the educational background of the provincial party secretary has a positive moderating effect on the reduction of environmental pollution through the establishment of AIIDEZ. Provincial party secretaries with a higher level of education are generally more sensitive and responsive to pilot reforms and capable of implementing more effective environmental protection policies while promoting technological innovation. This suggests that a higher level of education enhances the highest government leader\u0026rsquo;s understanding and ability to execute AIIDEZ policies, thereby intensifying the positive effects of AIIDEZ on environmental pollution governance.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eHeterogeneity analysis\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003e\u003cem\u003eM: Fiscal Autonomy\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003e\u003cem\u003eM: Highest Education of the Provincial Party Secretary\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eM \u0026times;Treat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.187***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.127*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.015**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.023*\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(0.064)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.069)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.007)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.013)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eTreat\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.020\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.009\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.122**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.157***\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(0.043)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.041)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.059)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.045)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eM\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.027\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.210\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.007\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.003\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e(0.185)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e(0.210)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e(0.008)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e(0.006)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEconomic controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMeteorology controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEnvironmental pollution controls\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eYear FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eProvince FE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithin R\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.127\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.314\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.108\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.311\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eObservation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e360\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNotes\u003c/em\u003e: Standard errors in parentheses are clustered at province level, \u003csup\u003e*\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.10, \u003csup\u003e**\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u003csup\u003e***\u003c/sup\u003e \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"5. Concluding remarks","content":"\u003cp\u003eThis study delves into the impact of the establishment of China\u0026rsquo;s AIIDEZ on government environmental performance, focusing on the key role of government in environmental governance. Through an empirical analysis, we find that the establishment of AIIDEZ significantly enhances government environmental performance by promoting green innovation and strengthening environmental regulations. However, when local governments have higher fiscal autonomy, they are more likely to prioritize economic growth, which weakens their enthusiasm for environmental governance. In addition, government leaders with higher levels of education are more sensitive and proactive in responding to policy reforms, which allows them to drive environmental governance more effectively and improve environmental performance.\u003c/p\u003e\u003cp\u003eThis study makes several contributions to the existing literature. While prior research on environmental governance has largely focused on corporate environmental responsibility (Ambec \u0026amp; Lanoie, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Porter \u0026amp; Linde, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e1995\u003c/span\u003e) and the role of technological innovation in firm-level environmental performance (Ghisetti \u0026amp; Rennings, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Horbach et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), our study shifts the analytical lens to the government level. We align with and extend the literature on technology-driven public governance (Margetts \u0026amp; Dunleavy, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Meijer \u0026amp; Bol\u0026iacute;var, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) by demonstrating how AI can enhance regulatory effectiveness and governance capabilities, ultimately improving environmental outcomes. Additionally, the findings contribute to the broader discourse on principal-agent problems in environmental policy implementation. Previous studies have highlighted the challenges local governments face in aligning national environmental goals with local economic priorities (Kostka \u0026amp; Hobbs, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Rhodes, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Our study provides empirical evidence that fiscal autonomy exacerbates these tensions, limiting the effectiveness of AI-driven environmental governance. This insight adds depth to existing research on decentralized governance and environmental policy execution (Berardo \u0026amp; Lubell, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Oates, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), suggesting that AI adoption alone is insufficient without appropriate institutional constraints and incentives.\u003c/p\u003e\u003cp\u003eBy framing AI as a catalyst for governance capacity enhancement, this study refines theoretical understandings of digital government and environmental performance. While digital transformation has been widely explored in public administration (Dunleavy, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Lips, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), its implications for environmental governance, particularly in emerging economies, remain underexamined. Our findings demonstrate that AI functions as both a policy instrument and an enforcement mechanism, bridging the gap between technological advancements and sustainable governance. From a policy perspective, this study highlights the necessity of complementary structural reforms to maximize AI\u0026rsquo;s governance potential. Given that fiscal autonomy may dilute environmental performance incentives, policymakers should consider environmental performance-linked fiscal transfers and AI-driven compliance monitoring to mitigate the trade-offs between economic growth and ecological sustainability. Furthermore, the observed correlation between leaders\u0026rsquo; education levels and effective AI-driven environmental governance underscores the need for targeted investments in administrative capacity-building and leadership training.\u003c/p\u003e\u003cp\u003eFuture research should systematically examine the heterogeneous impacts of AIIDEZ across varying economic and policy contexts, identifying key determinants that shape AI\u0026rsquo;s efficacy in environmental governance. Beyond AI, the exploration of emerging technologies such as blockchain, big data, and the Internet of Things (IoT) could expand the theoretical and empirical understanding of technology-enabled governance frameworks. Moreover, the institutional mechanisms governing AI\u0026rsquo;s integration into environmental policy demand rigorous scrutiny, particularly concerning transparency, accountability, and ethical considerations in algorithmic decision-making. Addressing these dimensions will not only refine governance models but also enhance the strategic deployment of AI and related technologies in advancing sustainable environmental policies.\u003c/p\u003e\u003cp\u003eIn conclusion, This study offers a novel perspective on AI\u0026rsquo;s application in governmental environmental performance, shifting the discourse from corporate environmental responsibility to state-led ecological governance. By contributing to the literature on environmental governance, digital public administration, and policy implementation, it advances theoretical and empirical understandings of technology-driven governance. As emerging economies grapple with the dual imperatives of technological innovation and environmental sustainability, our findings offer vital insights for policymakers and scholars alike. This study lays a robust foundation for future research at the intersection of AI, public governance, and environmental policy, underscoring the imperative for an interdisciplinary approach to sustainable development.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDisclosure statement of conflicts of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no conflict of interest.\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eNo funding was received for conducting this study.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eZ.L. designed the study, collected and analyzed the data, and wrote the initial draft of the manuscript. H.L. supervised the study, provided critical feedback, and revised the manuscript. Both authors reviewed and approved the final version of the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAcemoglu D, Restrepo P (2018) Artificial intelligence, automation, and work. The economics of artificial intelligence: An agenda. University of Chicago Press, pp 197\u0026ndash;236\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmbec S, Lanoie P (2008) Does it pay to be green? A systematic overview. Acad Manage Perspect, 45\u0026ndash;62\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBerardo R, Lubell M (2019) The ecology of games as a theory of polycentricity: Recent advances and future challenges. Policy Stud J 47(1):6\u0026ndash;26\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCao G, Weng X, Xu M, Zhou L-A (2023) Hybrid Contracts, Multitasking, and Incentives: Theory and Evidence from China's Air Pollution Controls. \u003cem\u003eMultitasking, and Incentives: Theory and Evidence from China's Air Pollution Controls (May 25, 2023)\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen J, Fleisher BM (1996) Regional income inequality and economic growth in China. J Comp Econ 22(2):141\u0026ndash;164\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen J, Huang C (2021) Policy Reinvention and Diffusion: Evidence from Chinese Provincial Governments. J Chin Polit Sci 26(4):723\u0026ndash;741\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDell M (2010) The Persistent Effects of Peru's Mining Mita. Econometrica 78(6):1863\u0026ndash;1903\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDong X, Zhou N, Zhao X, Yang S (2025) The impact of artificial intelligence on corporate green innovation: Can increasing quantity and improving quality go hand in hand? J Environ Manage 376:124439\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDunleavy P (2006) Digital era governance: IT corporations, the state, and e-government. Oxford University Press\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhisetti C, Rennings K (2014) Environmental innovations and profitability: How does it pay to be green? An empirical analysis on the German innovation survey. J Clean Prod 75:106\u0026ndash;117\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhobakhloo M, Fathi M (2021) Industry 4.0 and opportunities for energy sustainability. J Clean Prod 295:126427\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreenstone M, Hanna R (2014) Environmental regulations, air and water pollution, and infant mortality in India. Am Econ Rev 104(10):3038\u0026ndash;3072\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreenstone M, He G, Jia R, Liu T (2022) Can technology solve the principal-agent problem? Evidence from China\u0026rsquo;s war on air pollution. Am Economic Review: Insights 4(1):54\u0026ndash;70\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreenstone M, He G, Li S, Zou EY (2021) China\u0026rsquo;s war on pollution: Evidence from the first 5 years. Rev Environ Econ Policy 15(2):281\u0026ndash;299\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuo Q, Peng Y, Luo K (2025) The impact of artificial intelligence on energy environmental performance: Empirical evidence from cities in China. Energy Econ 141:108136\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe Q, Wei F, Zhang K, Zhong R, Kong F, Qi Y (2024) Fiscal decentralization, leader localization, and reduction of pollution and carbon emissions\u0026mdash;empirical evidence from China's fiscal province-managing-county reform. J Environ Manage 360:121175\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHong T, Yu NN, Mao ZG, Zhang SH (2021) Government-driven urbanisation and its impact on regional economic growth in China. Cities, \u003cem\u003e117\u003c/em\u003e, Article 103299.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHorbach J, Rammer C, Rennings K (2012) Determinants of eco-innovations by type of environmental impact\u0026mdash;The role of regulatory push/pull, technology push and market pull. Ecol Econ 78:112\u0026ndash;122\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHu S, Wang AL, Du KR (2023) Environmental tax reform and greenwashing: Evidence from Chinese listed companies. Energy Econ 124:106873\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHussain M, Yang S, Maqsood US, Zahid RA (2024) Tapping into the green potential: The power of artificial intelligence adoption in corporate green innovation drive. Bus Strategy Environ 33(5):4375\u0026ndash;4396\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJacobson LS, Lalonde RJ, Sullivan DG (1993) EARNINGS LOSSES OF DISPLACED WORKERS. Am Econ Rev 83(4):685\u0026ndash;709\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJia R, Nie H (2017) Decentralization, collusion, and coal mine deaths. Rev Econ Stat 99(1):105\u0026ndash;118\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJiang X, Li G, Fu W (2021) Government environmental governance, structural adjustment and air quality: A quasi-natural experiment based on the Three-year Action Plan to Win the Blue Sky Defense War. J Environ Manage 277:111470\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJin CF, Monfort A, Chen F, Xia N, Wu B (2024) Institutional investor ESG activism and corporate green innovation against climate change: Exploring differences between digital and non-digital firms. Technol Forecast Soc Chang, \u003cem\u003e200\u003c/em\u003e, Article 123129.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKanabkaew T, Mekbungwan P, Raksakietisak S, Kanchanasut K (2019) Detection of PM2. 5 plume movement from IoT ground level monitoring data. Environ Pollut 252:543\u0026ndash;552\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKluczek A, Zeglen P, Matus\u0026iacute;kov\u0026aacute; D (2021) The Use of Prospect Theory for Energy Sustainable Industry 4.0. Energies, \u003cem\u003e14\u003c/em\u003e(22), Article 7694.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKostka G, Hobbs W (2012) Local energy efficiency policy implementation in China: bridging the gap between national priorities and local interests. China Q 211:765\u0026ndash;785\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKostka G, Nahm J (2017) Central\u0026ndash;local relations: Recentralization and environmental governance in China. China Q 231:567\u0026ndash;582\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKou P, Han Y, Qi XY (2022) The operational mechanism and effectiveness of China's central environmental protection inspection: Evidence from air pollution. Socio-Economic Plann Sci 81:101215\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi HY, Lu ZQ, Zhang ZP, Tanasescu C (2025) How does artificial intelligence affect manufacturing firms' energy intensity? Energy Econ, \u003cem\u003e141\u003c/em\u003e, Article 108109.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi K, Lin B (2017) Economic growth model, structural transformation, and green productivity in China. Appl Energy 187:489\u0026ndash;500\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiang JQ, Langbein L (2015) Performance Management, High-Powered Incentives, and Environmental Policies in China. Int Public Manage J 18(3):346\u0026ndash;385\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLips M (2019) Digital government: Managing public sector reform in the digital era. Routledge\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu WJ, Liu CX, Luo J, Liu F (2024) How does digital transformation promote total factor productivity? Strategy, technology, and application. Manag Decis Econ 45(5):2739\u0026ndash;2750\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu Y, Wang B, Song Z (2025) Promoting or inhibiting: The impact of artificial intelligence application on corporate environmental performance. Int Rev Financial Anal 97:103872\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu ZF, Liu MD, Xia F, Zhang B (2024) Beyond technology: reducing environmental information asymmetry by centralizing environmental monitoring. Int Public Manage J 27(6):853\u0026ndash;873\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu J, Li B, Li H, Zhang XD (2020) Characteristics, exchange experience, and environmental efficiency of mayors: Evidence from 273 prefecture-level cities in China. J Environ Manage 255:109916\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLuo Z, Li H (2018) The impact of Atmosphere Ten Articles policy on air quality in China. China Ind Econ 9:136\u0026ndash;154\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMargetts H, Dunleavy P (2013) The second wave of digital-era governance: a quasi-paradigm for government on the Web. \u003cem\u003ePhilosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences\u003c/em\u003e, \u003cem\u003e371\u003c/em\u003e(1987), 20120382\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeijer A, Bol\u0026iacute;var MPR (2016) Governing the smart city: a review of the literature on smart urban governance. Int Rev Admin Sci 82(2):392\u0026ndash;408\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOates WE (2002) A reconsideration of environmental federalism. Recent advances in environmental economics. Edward Elgar Publishing\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePorter ME, Linde Cvd (1995) Toward a new conception of the environment-competitiveness relationship. J Economic Perspect 9(4):97\u0026ndash;118\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRazzaq A, Sharif A, Ozturk I, Yang XD (2023) Central inspections of environmental protection and transition for low-carbon Chinese cities: Policy intervention and mechanism analysis. Energy Econ 124:106859\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRhodes RAW (2018) Control and power in central-local government relations. Routledge\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSun YP, Guan WM, Cao YN, Bao Q (2022) Role of green finance policy in renewable energy deployment for carbon Evidence from China. Renewable Energy 197:643\u0026ndash;653\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTan YT, Mao XQ (2021) Assessment of the policy effectiveness of Central Inspections of Environmental Protection on improving air quality in China. J Clean Prod 288:125100\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang A (2022) Chinese Environmental Law. China Q 249:282\u0026ndash;284\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang A, Luo K, Nie Y (2024) Can artificial intelligence improve enterprise environmental performance: Evidence from China. J Environ Manage 370:123079\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang LK, Ma PP, Song Y, Zhang M (2023) How does environmental tax affect enterprises' total factor productivity? Evidence from the reform of environmental fee-to-tax in China. J Clean Prod, \u003cem\u003e413\u003c/em\u003e, Article 137441.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang Z, Zhang T, Ren X, Shi Y (2024) AI adoption rate and corporate green innovation efficiency: evidence from Chinese energy companies. Energy Econ 132:107499\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWei J, Li ZQ, Cribb M, Huang W, Xue WH, Sun L, Guo JP, Peng YR, Li J, Lyapustin A, Liu L, Wu H, Song YM (2020) Improved 1 km resolution PM2.5 estimates across China using enhanced space-time extremely randomized trees. Atmos Chem Phys 20(6):3273\u0026ndash;3289\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWei J, Li ZQ, Lyapustin A, Sun L, Peng YR, Xue WH, Su TN, Cribb M (2021) Reconstructing 1-km-resolution high-quality PM2.5 data records from 2000 to 2018 in China: spatiotemporal variations and policy implications. Remote Sens Environ 252:112136\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu RX, Hu P (2019) Does the Miracle Drug of Environmental Governance Really Improve Air Quality? Evidence from China's System of Central Environmental Protection Inspections. Int J Environ Res Public Health, \u003cem\u003e16\u003c/em\u003e(5), Article 850.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu H, Li Y, Lin W, Li Y (2024) Government fiscal decentralization and haze and carbon reduction: Evidence from the fiscal Province-Managing-County reform. Environ Res 252:119020\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang C-H (2022) How artificial intelligence technology affects productivity and employment: firm-level evidence from Taiwan. Res Policy 51(6):104536\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang L, Lin Y, Wang J, Peng F (2024) Achieving Air Pollution Control Targets with Technology\u0026ndash;Aided Monitoring: Better Enforcement or Localized Efforts? Am Economic Journal: Economic Policy 16(4):280\u0026ndash;315\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang M, Chou S-Y (2018) The impact of environmental regulation on fetal health: evidence from the shutdown of a coal-fired power plant located upwind of New Jersey. J Environ Econ Manag 90:269\u0026ndash;293\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYao J, Zhang K, Guo L, Feng X (2024) How does artificial intelligence improve firm productivity? Based on the perspective of labor skill structure adjustment. J Manage World 40(2):101\u0026ndash;116\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYi F, Niu ZH, Zhong SB (2025) Does information and communication technology reduce carbon emissions in China? Evidence from the quasi-natural experiment of the Broadband China pilot policy. J Environ Planning Manage 68(1):1\u0026ndash;27\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYing Y, Cui XY, Jin SY (2023) Artificial Intelligence and Green Total Factor Productivity: The Moderating Effect of Slack Resources. Systems, \u003cem\u003e11\u003c/em\u003e(7), Article 356.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang DY (2022) Green financial system regulation shock and greenwashing behaviors: Evidence from Chinese firms. Energy Econ 111:106064\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang GL (2023) Regulatory-driven corporate greenwashing: Evidence from ?low-carbon city? pilot policy in China. Pac-Basin Financ J, \u003cem\u003e78\u003c/em\u003e, Article 101951.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang K, Kou ZX, Zhu PH, Qian XY, Yang YZ (2025) How does AI affect urban carbon emissions? Quasi-experimental evidence from China's AI innovation and development pilot zones. Econ Anal Policy 85:426\u0026ndash;447\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang P, Wu JN (2018) Impact of mandatory targets on PM2.5 concentration control in Chinese cities. J Clean Prod 197:323\u0026ndash;331\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhou W, Zhuang Y, Chen Y (2024) How does artificial intelligence affect pollutant emissions by improving energy efficiency and developing green technology. Energy Econ 131:107355\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhu XF, Wang Y (2024) Credible signaling to promote local compliance: Evidence from China's multiwave inspection of environmental protection. Public Adm\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhuang MX (2024) Supervising Local Cadres in China: The Quest for Authoritarian Accountability. Politics Soc 52(3):452\u0026ndash;485\u003c/span\u003e\u003c/li\u003e\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, Government Environmental Performance, Fiscal Autonomy, Leader Education Level","lastPublishedDoi":"10.21203/rs.3.rs-7322815/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7322815/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAs environmental governance becomes a central issue of global concern, how to use emerging technologies to improve government environmental performance has become an important area of research. This study examines the impact of the establishment of China\u0026rsquo;s Artificial Intelligence (AI) Innovation Pilot Zones on local government environmental performance, with a focus on the key role of government in environmental governance. Based on data from three batches of AI innovation pilot zones between 2019 and 2021, we find that AI pilot zones significantly enhance local government environmental performance by promoting green innovation and strengthening environmental regulations. However, higher fiscal autonomy to some extent weakens the actual effectiveness of environmental policies, while leaders with higher education levels are more effective in implementing environmental policies. This study expands the understanding of technology-driven environmental governance and provides new theoretical perspectives for environmental policy in emerging economies, offering empirical support for the role of technology in empowering public governance.\u003c/p\u003e","manuscriptTitle":"Technological Transformation of Environmental Governance: Evidence from a Quasi-Natural Experiment in China","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-01 12:27:25","doi":"10.21203/rs.3.rs-7322815/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":"8728f1a8-cb08-4c49-b78d-6869bf143e89","owner":[],"postedDate":"December 1st, 2025","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-05-13T16:41:39+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-04T18:13:56+00:00","index":112,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":58775174,"name":"Earth and environmental sciences/Environmental sciences"},{"id":58775175,"name":"Earth and environmental sciences/Environmental social sciences"},{"id":58775176,"name":"Social science/Environmental studies"}],"tags":[],"updatedAt":"2026-05-13T16:55:30+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-01 12:27:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7322815","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7322815","identity":"rs-7322815","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}