Non-linear research on artificial intelligence empowering green economic efficiency under integrated governance framework

Artificial intelligence (AI) plays a pivotal role in the development of the green economy. This paper examines the impact of artificial intelligence (AI) on green economic efficiency (GEE) using panel data from 30 provinces in China spanning 2011–2020. A multiple linear regression model, alongside various endogeneity and robustness tests, is applied to ensure reliable findings. The empirical results indicate that AI significantly enhances GEE. However, the marginal effect of AI on GEE is influenced by different governance approaches. In terms of policy governance, excessive market-based environmental regulation (MER) diminishes the marginal impact of AI, while stronger administrative-command environmental regulations (CER) and informal environmental regulations (IER) amplify it. Regarding technological governance, substantive green technological innovations (SUG) reduce AI's marginal effect, whereas symbolic green technological innovations (SYG) may increase it. Notably, the threshold effect of SUG surpasses that of SYG. In legal governance, both administrative and judicial intellectual property protections reduce the marginal effect of AI, though administrative protection (AIP) exhibits a more significant threshold effect than judicial protection (JIP). These findings offer practical insights for optimizing governance strategies to maximize AI's role in promoting GEE. These insights highlight the need for balanced governance to maximize AI's role in sustainable development. Policymakers should tailor regulations and encourage regional collaboration to harness AI's spatial spillover effects. Enterprises can leverage AI-driven innovations to align growth with ecological goals, fostering coordinated green development.

1 Introduction

In recent decades, China's economy has achieved remarkable growth, positioning it as one of the world's fastest-growing nations, often called the “Chinese economic miracle.” However, this rapid economic expansion has come at a considerable cost to the country's resources and environment, leading to severe pollution and resource depletion (Rasheed et al., 2024a). These environmental challenges now act as significant barriers to the sustainable development of the Chinese economy and society (Gao et al., 2024), with potential long-term detrimental effects. As illustrated in Figure 1, Chinese energy consumption dramatically surpasses the other major economies, positioning it as the globe's largest energy-consuming country. In the past 2023, China's energy usage was about 1.8 times as much as the United States, 4.3 times as much as India, 10 times as much as Japan, and 15 times as much as Germany. The Chinese government realizes that it is not desirable to judge heroes by GDP, and it is necessary to abandon the “speed syndrome,” a developmental mindset in which speed produces political results. The Chinese government has called for attention to the problems of environmental pollution and waste of resources in the meantime of economic development, and to adhere to green growth, raise the green economic efficiency (GEE), and get rid of the rough-and-tumble mode of development. It advocates for integrated measures to reduce carbon emissions and pollution, expand green spaces, foster ecological prioritization, conservation, and green low-carbon growth, and push forward comprehensive green transformation in economic development (Lin and Zhou, 2022).

Figure 1

Figure 1. Energy consumption in the world's major economies.

GEE represents a measure that integrates economic growth, resource consumption, and environmental pollution. This indicator assesses the economic efficiency of a country or region while accounting for environmental costs, thereby providing a more comprehensive and accurate reflection of the impact of economic activities (Liu and Dong, 2021). With the advent of AI, China must utilize AI's power to achieve sustainable green development and improve GEE. AI, a seminal “general-purpose technology” following the internet, is pivotal in driving the fourth industrial revolution. AI, leveraging technologies such as deep learning, pattern recognition, and autonomous decision-making, not only supplants traditional technologies but also facilitates rapid evolutionary and iterative advancements through its integration with various application sectors (Abulibdeh et al., 2024; Konya and Nematzadeh, 2024). AI fosters the development of more sustainable and cleaner technologies, decreasing consumption in production processes (Akter et al., 2024; Olawade et al., 2024; Konya and Nematzadeh, 2024). AI has notably enhanced the energy structure, with improvements in energy efficiency identified as the primary factor contributing to the deceleration of CO₂ emissions (Rasheed et al., 2024c). Additionally, AI bolsters green development awareness, ultimately steering societal shifts toward energy-saving and low-carbon behaviors (Nahar, 2024). AI is also seen as one of the tools to tackle air pollution effectively (Shaamala et al., 2024; Cai et al., 2024; Adnan et al., 2024). As a result, AI has infused the green economy with new momentum and emerged as a crucial driver of GEE improvements (Bibri et al., 2023).

Existing studies have established the significant role of AI in advancing green development, such as its contributions to technological innovation, energy efficiency, and carbon emission reduction. However, two critical gaps persist in the literature. First, while AI's general impact on green development has been explored, limited attention has been given to its direct and non-linear effects on GEE, particularly in the context of varying governance frameworks. Second, existing research often isolates governance dimensions rather than adopting an integrated approach that collectively considers policy, technology, and legal mechanisms. This fragmentation overlooks the threshold effects among governance types. To address these gaps, this study uniquely integrates AI with GEE within a comprehensive governance framework, offering a novel perspective on the heterogeneous and threshold effects across governance types. By employing panel data from 30 Chinese provinces (2011–2020), this research provides robust empirical evidence that enriches the understanding of AI's nuanced role in sustainable development.

This paper makes three significant theoretical contributions. Firstly, this paper uniquely integrates AI with GEE within a unified research framework concerning research subjects. Secondly, regarding mechanism pathways, the paper seeks to elucidate the non-linear effects of AI on GEE, contextualized within an overarching governance framework encompassing policy, technology, and law. The AI-empowered GEE is different under the roles of varying kinds of environmental regulation, green technology innovation, and intellectual property protection. Thirdly, regarding research effects, this paper considers the different geographic locations, technological levels, degree of openness, and differences in the effectiveness of AI-empowered GEE.

The structure of this paper is as follows: Section 2, “Literature Review,” summarizes relevant studies. Section 3 presents the theoretical analysis and proposes research hypotheses. Section 4 describes the research design, including the model, variable measurements, and data sources. Section 5 provides an empirical study of AI's direct impact on GEE. Section 6 examines the non-linear effect of AI on GEE under the influence of comprehensive governance. Finally, Section 7 concludes the study, summarizes key findings, and offers recommendations.

2 Literature review

A thorough literature review, conducted across the ScienceDirect and Web of Science databases up to July 2024, uncovers an extensive corpus of scholarly work focused on economic efficiency and the green economy paradigm. Nonetheless, a significant absence of academic discourse is evident regarding the specific concept of GEE. Moreover, the direct impact of AI on bolstering GEE has been explored in an exceedingly sparse number of studies, as illustrated in Figure 2.

Figure 2

Figure 2. Research status of artificial intelligence empowering green economy efficiency.

This study integrates insights from multiple theoretical perspectives to construct a comprehensive governance framework. Policy governance draws on the Porter Hypothesis, which posits that appropriate environmental regulations can stimulate innovation and offset compliance costs (Porter and van der Linde, 1995). Technological governance builds upon the theory of technological progress, emphasizing the transformative role of green innovations in improving resource efficiency (Dao et al., 2024). Legal governance leverages institutional theories, highlighting how intellectual property protection can foster innovation while potentially creating barriers to knowledge diffusion (DiMaggio and Powell, 1983). Unlike prior studies focusing on isolated dimensions, this study synthesizes these theoretical approaches into an integrated framework, providing a holistic view of how governance moderates AI's influence on GEE.

Some scholars have already paid attention to AI's enormous energy and explored whether AI can contribute to green development. AI technologies significantly foster the development of green technologies and contribute to developing a green economy (Zhao et al., 2022). AI positively influences green economic transformation through its varied effects on internal factors and markedly reduces carbon intensity in conjunction with natural resource markets (Feng et al., 2024). Notably, industrial robots decrease pollutant emission intensity, sharply reduce energy efficiency, and have a noticeable negative spatial spillover effect on carbon emission efficiency (Zhou et al., 2024a,b). AI supports the sustainable development of urban environments through real-time monitoring and data-driven approaches, facilitating the achievement of green development goals (Bibri et al., 2024).

Several scholars have discussed the impact of technological advances on GEE. They highlighted significant contributions from innovations in renewable energy technologies and the digital economy, which boosted urban GEE (Chen and Yang, 2024; Huang et al., 2023). Technological innovations densely concentrated in specific areas enhance GEE, while digital finance drives urban GEE via green technology innovations (Liu and Dong, 2021; Li and Xu, 2023). In addition, Yang et al. (2021b) argues that technological progress contributes to the decoupling of economic growth from CO₂ emissions and that Internet development can curb haze pollution through technological innovations, confirming technological advances' impact on GEE.

Others have explored the impact of other socio-economic factors on GEE. Some pointed out that human capital mismatches considerably detract from GEE, while openness and environmental regulation exhibit significant U-shaped non-linear effects (Shuai and Fan, 2020). Digital transformation has notably increased GEE by 2.6%, and economic resilience interacts beneficially with GEE (Lv and Chen, 2024; Wang and Zhou, 2024). These studies collectively provide a robust theoretical foundation and logical starting point for the in-depth exploration in this paper. Nevertheless, there is a notable gap in discussions concerning the immediate effect of AI on GEE, the mechanisms through which AI enhances GEE, and the potential heterogeneity in AI's effects.

In summary, existing research on AI's role in promoting green sustainable development and the factors influencing GEE has yielded significant findings. However, few studies have specifically examined AI's direct effects on GEE. Additionally, the potential role of integrated governance in the AI-empowered GEE relationship remains largely unexplored. To address these gaps, this paper investigates the impact of AI on GEE and the role of integrated governance, thereby extending previous research.

3 Theoretical analysis and research hypothesis

3.1 The direct impact of artificial intelligence on green economic efficiency

On the one hand, it aligns with traditional theories of technological progress, positioning AI as the apex of various advanced information technologies. AI shares characteristics with other information technologies yet possesses unique technical features that distinguish it from its predecessors (Ahmed et al., 2022). Secondly, AI significantly enhances production process efficiency through automation and intelligent systems, reducing waste and optimizing resource allocation, lowering energy and raw material consumption per output unit, thereby improving GEE (Tian et al., 2023). Thirdly, as a pioneering technology, AI supports the development of clean energy technologies, fostering innovation in the green economy (Li et al., 2024). Fourthly, AI's application in energy management optimizes energy consumption patterns, ensures precise energy use, minimizes unnecessary waste, and enhances overall energy efficiency. Within the supply chain, AI predicts demand, optimizes inventory management, reduces transportation costs, and lowers the carbon footprint (Rasheed et al., 2024b).

On the other hand, from an economic system efficiency optimization perspective, the widespread adoption of AI blurs the boundaries between different sectors' economic activities, strengthens horizontal and vertical industry linkages, and promotes industrial agglomeration and collaborative development, especially among clean, high-tech industries (Yang et al., 2021a, 2022). AI has reduced outdated production capacity and pollutant emissions, which positively influenced the green development of the region (Liu and Zhang, 2021). Moreover, AI's broad application establishes intelligent economic circulation channels at the macro level, considerably improving production and resource utilization efficiencies. These efficient channels mitigate the adverse effects of external market segmentation and other inhibitory factors, facilitating economic efficiency advances. AI supports balanced development across economic growth, resource conservation, and environmental protection. Consequently, this study proposes the hypothesis:

H1: AI may directly enhance GEE.

3.2 The non-linear effect of artificial intelligence on green economic efficiency under integrated governance

Integrated governance involves multiple stakeholders, including governments, enterprises, social organizations, and citizens, sharing resources and collaborating to address public affairs and maximize public interest. The integrated governance model breaks through the limitations of the traditional government's single-subject management and emphasizes multiple subjects' joint participation and synergy (Amsler, 2016). These stakeholders leverage their strengths to create complementary effects during the governance process. For instance, governments can augment GEE, driven by AI, by establishing mandatory regulations, supportive policies, and incentives such as tax benefits, R&D subsidies, and green credits. Social organizations and the public amplify their influence through media engagement, while companies invest in developing green technologies and AI. Judicial institutions safeguard the rights of innovators in the green economy by enacting and enforcing intellectual property laws. Collectively, these actions contribute to advancements in GEE facilitated by AI. Consequently, this paper explores how AI enhances GEE within the framework of collaborative governance involving policy, technology, and law, as shown in Figure 3.

Figure 3

Figure 3. Theoretical framework.

3.2.1 The non-linear impact of artificial intelligence on green economic efficiency under policy governance

With the prominence of resource and environmental problems, China has adopted a series of ecological regulatory instruments to alleviate the issues of resource depletion and environmental pollution. According to the theory of “Porter's hypothesis,” adopting specific environmental regulation measures can motivate the innovation vitality of corporations and Strengthen the competitiveness of products, thus offsetting or even exceeding the cost of compliance brought about by environmental regulations and facilitating the green economy. Therefore, examining the impact of environmental regulations on AI-empowered GEE is crucial for establishing a green economic system and achieving the “dual carbon” goals. Environmental regulation encompasses a comprehensive set of policies, laws, standards, and measures developed by governments and relevant agencies to protect the environment, reduce pollution, improve ecological conditions, and guide or restrict the actions of enterprises, individuals, and other societal actors. Its primary goal is minimizing negative environmental impacts while supporting sustainable development and economic growth. The three main types of environmental regulation—market-based environmental regulation (MER), administrative-command environmental regulations (CER), and informal environmental regulation (IER)—represent the approaches employed by governments, markets, and society. Together, these regulatory dimensions provide a holistic framework for assessing the intensity and effectiveness of environmental regulation through mandatory control measures, economic guidance tools, and social constraints (Song et al., 2024). This study categorizes environmental regulations into market-based, administrative-command, and informal for analytical purposes (Wang et al., 2022, 2023a).

MER is the government's way of guiding enterprises to green innovation by establishing special funds for research and development, environmental protection, technological reform, tax breaks, loan preferences, financial subsidies, and other means. MER shapes corporate environmental strategies, steering businesses from traditional end-of-pipe pollution control toward digital and cleaner production models, thereby enhancing GEE (Yang et al., 2023; Kazemzadeh et al., 2023). However, overly stringent MER may escalate operational costs for businesses, diminishing their inclination to invest in AI. Companies might prioritize short-term investments with immediate cost benefits over long-term sustainable growth, thus undermining commitment to green transformation (Koengkan et al., 2024).

CER, enacted by government bodies, sets precise environmental standards and mandates for compliance, compelling technological innovation, leading to more accurate investments in AI and green technologies, and the firms to explore new green products and services to enhance GEE (Naeem et al., 2024; Chen et al., 2021). Nevertheless, suppose CER is loosely enforced, and penalties for non-compliance are minimal. In that case, firms may be inclined to ignore regulations to cut costs, potentially exacerbating pollution and resource exploitation and worsening environmental challenges.

IER refers to norms and behaviors that are not legally enforceable but which can influence the environmental actions of enterprises and individuals. These regulations typically originate from non-governmental sources such as social culture, public awareness, civil society organizations, and the media. The public increases demand for environmentally friendly products and actively participates in environmental management by reporting polluting behaviors to the media and authorities. This will curb corporate misbehavior and enhance corporate social responsibility. It motivates enterprises to adopt AI to analyze consumer demand accurately, optimize production processes to reduce environmental pollution and improve GEE (Zhang et al., 2023, 2022). Therefore, this paper formulates the following hypotheses:

H2a: Excessive MER weakens AI-empowered GEE.

H2b: CER strengthens AI-empowered GEE.

H2c: IER enhances AI-empowered GEE.

3.2.2 The non-linear impact of artificial intelligence on green economic efficiency under technical governance

As green technology innovation progresses, the enhancement of GEE by AI exhibits variability. Drawing from studies by Lian et al. (2022), it is evident that companies undertake both substantive green technology innovations (SUG) and symbolic green technology innovations (SYG) driven by varying motivations. SUG, characterized by long development cycles, significant challenges, and substantial investments, delivers high degrees of innovation and technical content, epitomizing high-quality green innovation activities. However, they are subject to cumulative and threshold effects; excessive investment in these innovations without timely monetization can escalate production costs, potentially restricting other non-green R&D activities and diminishing a company's overall competitiveness and profitability, thereby reducing the potential enhancement effect of AI on GEE (Fang et al., 2022). SYG involves secondary innovations that extend functionalities and improve technology without altering the fundamental technological principles of products. These innovations, which have shorter development cycles and require lesser investment, can be quickly integrated into corporate production and development, offering practical benefits. They primarily align with government environmental regulatory strategies or seek advantages through the “quantity” and “speed” of innovation. Requiring less investment and being more strategically driven, SYG allows companies to rapidly respond to market or policy pressures without significantly impacting other non-green R&D efforts. SYG enhances publicity, improves competitiveness, increases profits, and promotes further investment in AI and green technologies, thereby strengthening AI's role in enhancing GEE. Compared to the more superficial SYG, SUG stands for a profound technological breakthrough that has had a remarkable impact on optimizing production processes, improving resource efficiency, and reducing environmental pollution (Chen et al., 2024b). The influence of SUG on GEE facilitated by AI is markedly significant. Therefore, this paper formulates the following hypotheses:

H3a: Excessive SUG weakens AI-empowered GEE.

H3b: SYG strengthens AI-empowered GEE.

H3c: The threshold effect of SUG is more potent than SYG's.

3.2.3 The non-linear impact of artificial intelligence on green economic efficiency under legal governance

Intellectual property is defined as a monopolistic property right granted to creators over their intellectual creations. Intellectual property protection ensures that rights holders can fully operate these rights. Due to the externality characteristics of green innovation activities, it is difficult for innovation subjects to avoid imitation or even plagiarism by other competitors, jeopardizing innovation subjects' vested interests and weakening their willingness to carry out green innovation activities (Hall and Helmers, 2013). By reducing the externality of green innovation, intellectual property protection helps secure the monopoly profits of innovators (Song et al., 2022). However, excessive protection of these monopoly rights can discourage the sharing of technological achievements, potentially diminishing the monopolist's profits (Song and Chen, 2023). Based on the theory of “economic man,” rights holders are generally disinclined to share under these circumstances. Excessive intellectual property protection reinforces the monopoly position of the rights holder, thereby impeding the diffusion of technology and weakening the incentives of other entities to research, develop, and innovate. Therefore, excessive administrative intellectual property protection (AIP) and judicial intellectual property protection (JIP) can restrict GEE empowerment by AI. AIP, typically managed by government agencies, directly influences corporate decisions regarding technological and green innovation. JIP, handled by courts, addresses complex issues of technological infringement and interprets intellectual property laws. Given their generally more independent nature, judicial institutions tend to have a less direct impact on business operations. The adverse effects of excessive AIP, imposed by governments, are generally more significant than those resulting from excessive JIP. Therefore, this paper formulates the following hypotheses:

H4a: Excessive AIP weakens AI-empowered GEE.

H4b: Excessive JIP weakens AI-empowered GEE.

H4c: The threshold effect of AIP is greater than that of JIP.

4 Research design

4.1 Econometric model

To substantiate the enhancement of GEE by AI, the following econometric model is constructed:

$\begin{array}{ll}Ge{e}_{it}=\alpha +\beta A{I}_{it}+\varsigma {\displaystyle \sum Control+u_t+{\lambda }_i+{ϵ}_{it}}& (1)\end{array}$

In this model, Gee represents green economic efficiency, AI represents artificial intelligence, Control is control variable, u_t and λ_i represent time effects and individual effects, and ϵ is the random disturbance term.

4.2 Variable descriptions

4.2.1 Dependent variable: green economic efficiency

The traditional Data Envelopment Analysis (DEA) model is widely used to evaluate the efficiency of decision-making units; however, it cannot account for the impact of undesirable outputs on efficiency and fails to differentiate among multiple efficient units (Tone, 2001). To address these limitations, Tone improved the DEA model by introducing a non-radial, non-angular super-efficiency DEA approach that incorporates undesirable outputs, known as the super-efficiency Slack-Based Measure (SBM) model based on undesirable outputs (Aparicio et al., 2017). This model integrates relaxation variables and resolves the issue of distinguishing efficiency values >1 in traditional SBM models, making it widely applicable in efficiency measurement (Li and Shi, 2014). This paper draws on the research of Zhou et al. (2018) to measure GEE by adopting the super-efficient SBM model, including non-expected outputs, selecting labor force, capital, electricity, water resources, and natural gas resources as the input indexes, and selecting the GDP to represent the expected outputs, and selecting the “three wastes” are chosen to represent the non-desired outputs (Table 1). The GEE is calculated as follows:

$\begin{array}{ll}Ge{e}^{*}=\frac{\frac{1}{m}{\displaystyle \sum _{i=1}^m}\frac{{\overline{x}}_i}{x_{ik}}}{\frac{1}{s_1+s_2}\left({\displaystyle \sum _{r=1}^{s_1}}\frac{{\overline{y}}_r^g}{y_{rk}^g}+{\displaystyle \sum _{r=1}^{s_2}}\frac{{\overline{y}}_r^b}{y_{rk}^b}\right)}& (2)\end{array}$

$\begin{array}{ll}S.t.\{\begin{array}{l}\overline{x}\ge {\displaystyle \sum _{j=1,j\ne k}^n}{x}_j{\lambda }_j\\ {\overline{y}}^g\le {\displaystyle \sum _{j=1,j\ne k}^n}{y}_j^g{\lambda }_j\\ {\overline{y}}^b\ge {\displaystyle \sum _{j=1,j\ne k}^n}{y}_j^b{\lambda }_j\\ \overline{x}\ge x_0,{\overline{y}}^g\le y_0^g,{\overline{y}}^b\ge y_0^b,{\overline{y}}^g\ge 0,\lambda \ge 0\end{array}& (3)\end{array}$

Table 1

Table 1. Green economic efficiency evaluation index system.

Where Gee^* denotes the value of GEE, n represents the number of the study area, k denotes the serial number of each province, m is the input of each province, s₁ and s₂ denotes the desired output and non-desired output of each province, respectively, x denotes the input indicator of each province for the period, y^g and y^b denote the desired output indicator and non-desired output indicator of each province during the period, respectively.

4.2.2 Core explanatory variable: artificial intelligence

Following the approach of Liu et al. (2020), this paper uses the number of industrial robots installed in each province to represent AI. The calculation method, shown in Equation 4, reflects the relationship between robot stock and labor supply, potentially offering a more effective measure of AI (Tao, 2024).

It is assumed that the distribution of industrial robots across Chinese provinces is uniform, meaning each province has the same industrial robot installation density. Thus, the industrial robot installation density = total national industrial robot installations/total national employment. Therefore, the industrial robot installations in each province = industrial robot installation density × employment in each province. The specific formula is as follows:

$\begin{array}{ll}A{I}_{it}=L_{it}\times \frac{Robo{t}_t}{L_t}& (4)\end{array}$

Equation 4, AI_it indicates the number of industrial robots installed in the i region during the t period; Robot_t indicates the total number of industrial robots installed nationally during the t period; L_it indicates the number of employed persons in the i region during the t period; L_t indicates the total number of employed persons nationally during the t period. To eliminate the effect of dimension, the natural logarithm is taken for the number of industrial robots installed in each province.

4.2.3 Control variables

This paper selects the following control variables to make the model estimation results more accurate: internet infrastructure conditions (IIS), expressed using the logarithm of number of internet broadband access ports; level of economic development (PGDP), expressed using the logarithm of per capita gross regional product; industrial structure (ISU), expressed using the region's share of the value added of the tertiary industry in the value added of the secondary sector; level of industrialization (IL), described by the ratio of the current year's industrial added value to the current year's GDP (Zhou et al., 2024b; Feng et al., 2024).

4.2.4 Policy governance: environmental regulation

MER: measured by the ratio of completed investment in industrial pollution treatment to industrial value added (Xu, 2024). This indicator reflects the financial incentives and cost pressures imposed on firms to adopt environmentally friendly practices, aligning with the Porter Hypothesis, which suggests that appropriate environmental regulations can stimulate innovation.

CER: measured by the logarithm of granted environmental administrative penalty cases (Zhang et al., 2024b). This captures the enforcement strength of environmental policies, directly reflecting the regulatory pressure imposed on enterprises to comply with environmental standards.

IER: constructed using entropy weighting of factors such as per capita disposable income (per capita disposable income of urban residents), educational levels (proportion of employees with college education or above), population density (year-end permanent population measure) and age Structure (measured by the proportion of population under 15 years old) (Shen et al., 2023). These components represent societal engagement in environmental governance, highlighting public awareness and the role of social norms in promoting green economy.

4.2.5 Technical governance: green technology innovation

Green patents intuitively reflect the extent and level of green technology innovation. Considering that granted patents have undergone authoritative certification, this paper uses the logarithm of granted green invention and utility patents to depict SUG and SYG (Lian et al., 2022; Zhang et al., 2024a).

4.2.6 Legal governance: intellectual property protection

The number of cases closed refers to the cases adjudicated and closed, reflecting the achievements and the capacity of judicial or administrative authorities to handle cases within a certain period. The number of cases filed may be influenced by selective filing, which has high uncertainty. Thus, the number of cases closed provides a more objective measure. This paper uses the logarithm of administrative closures of patent infringement cases to indicate the level of AIP and the logarithm of judicial closures of patent infringement cases to indicate the level of JIP (Li and Yuan, 2024).

4.3 Data sources

Considering the fact that Chinese officials did not publish the China Artificial Intelligence White Paper until 2015, the relative lack of data on AI before 2015, and the lag in the disclosure of data on industrial robots by the International Federation of Robotics (IFR), this paper collects the panel data of 30 provinces in mainland China (Missing data for Tibet, not included in statistics) from 2011 to 2020 as the research sample. The data come from the patent database of the China Intellectual Property Office, the National Bureau of Statistics, the China Statistical Yearbook, and statistical bulletins issued by provinces. All variables are standardized to improve estimation accuracy and eliminate the effects of dimensions. Table 2 presents the variable descriptive statistics.

Table 2

Table 2. Descriptive statistics results.

5 Direct effects of artificial intelligence on green economy efficiency

5.1 Benchmark regression analysis

Based on the econometric model (1), a fixed effects model was applied following the Hausman test, controlling time and individuals. Multiple regression analysis was then used to assess the effect of AI's empowerment on GEE. The relationship between AI and GEE under the dynamic effect of various variables is examined by gradually introducing control variables. The benchmark regression results are displayed in Table 3. The estimation results reported in Model (1)–Model (5) show that the regression coefficients and significance of the core explanatory variables and control variables did not change radically during the gradual approach to introducing the control variables. Considering control variables, for every 1% improvement in AI, GEE increases by about 0.281%, which is significant at the 1% level, indicating that AI can improve GEE and preliminarily verify H1.

Table 3

Table 3. Benchmark regression results.

5.2 Robustness tests

To ensure the robustness of the results of the benchmark regression, this paper conducts robustness tests from three aspects: winsorize, changing estimation methods, and replacing core variables, with results shown in Table 4. First, for winsorize, using the method of deleting 1% extreme samples, the results are shown in column (1) of Table 4; the coefficient of AI is 0.281 and passes the 1% significance level test. Second, this paper employs the system GMM regression model for re-estimation, with results displayed in column (2). The regression coefficient for AI is 0.512, which is significant at the 1% level. Third, by changing the independent variable to the number of AI companies as a substitute, the results in column (3) show an AI regression coefficient of 0.595, tested by the 1% significance level. These robustness test results are aligned with the benchmark regression, further verifying the positive influence of AI on GEE.

Table 4

Table 4. Robustness tests regression results.

5.3 Endogeneity tests

To address the endogeneity, this paper employs the two-stage instrumental variable least squares (IV-2SLS) method for testing. This paper uses the first lag of the dependent variable as an instrumental variable. The endogeneity test results in Table 5 indicate that the instrumental variable is significantly at the 1% level and passes the weak identification test and the non-identification test, confirming the effectiveness of the instrumental variable selection. According to column (2), the original hypothesis that AI facilitates GEE remains valid. The conclusion of the benchmark regression is robust and reliable, thereby verifying H1.

Table 5

Table 5. Endogeneity tests regression results.

5.4 Heterogeneity analysis

To verify the geographical heterogeneity of AI's impact on GEE, this study categorizes the samples based on economic levels into Eastern, Central, and Western regions; based on openness into coastal and non-coastal areas; and based on Technology Intensity into technology-intensive and non-technology-intensive regions for a heterogeneity analysis.

First, in the Eastern and Central regions, the coefficients of AI on GEE are 0.331 and 0.209, and by the test of significance at the 1% level, which is insignificant for the Western region. This indicates that AI contributes the most to GEE in the East, followed by the Center, and not significantly in the West. Next, in coastal areas, the coefficients of AI on GEE are 0.329 and 0.106, significant at the 1% level, suggesting that AI has a more substantial impact on GEE in coastal areas than in non-coastal areas. Lastly, in technology-intensive areas, the coefficient of AI on GEE is 0.440, significantly at the 1% level, while in non-technology-intensive areas, the coefficient is positive but insignificant. This reveals that AI fosters GEE in technology-intensive regions, but its impact is insignificant in non-technology-intensive areas.

5.5 Spatial effects test of artificial intelligence empowering green economic efficiency

AI not only promotes the efficient flow of technology and information, breaking through the constraints of time and space and compressing geographic limitations, impacting the economic system with long-term, extensive, and holistic effects; on the other hand, the rapid advancement of AI will unleash the tremendous energy of the “connected economy,” generating demonstration effects and economic linkage effects in geographically and economically adjacent areas through technology spillover, knowledge sharing, talent exchange, and regional coordination. Thus, from the standpoint of spatial spillover effects, further analyzing the role of AI-driven technology in enhancing GEE can provide valuable decision-making references and policy insights for authorities to comprehensively assess AI technology features, formulate forward-looking industrial policies, and create a new pattern of regional synergy optimization in GEE.

5.5.1 Moran's I test

To recognize spatial effects, the first step is to use the Moran's I index for testing spatial correlations (building an economic inverse distance spatial weight matrix W using the reciprocal of provincial capital distances, then integrating economic factors to construct an economic and geographical nested matrix), with results shown in Table 7. Moran's I index of AI is over 0 and passes the significance test in most years, indicating that AI has strikingly positive spatial dependence properties. Figure 4 illustrates the spatial distribution of AI for the years 2011, 2015, 2018, and 2020, consistently showing positive spatial correlation. In addition, the Moran'I index of GEE is over 0 in both matrices. Most years, it passes the significance test, indicating that GEE demonstrates a clear positive spatial dependence property.

Table 6

Table 6. Heterogeneity test results.

Table 7

Table 7. Moran's I indices for green economic efficiency and artificial intelligence.

Figure 4

Figure 4. Spatial distribution of artificial intelligence.

5.5.2 Spatial econometrics test

In this paper, we compare the fitness of SAR and SEM using the LM test and further use Wald and LR to test whether SDM can degrade to SAR and SEM. Table 8 shows that the LM lag and RLM lag tests in the geographic weighting matrix satisfy the 1% level of significance, the LM error defies the 5% level of significance, and the RLM error does not fulfill the significance test; the economic-geographic nesting matrix in which the LM lag test for significance at the 5% level, the RLM lag is tested for significance at the 1% level, the LM error and RLM error do not pass the significance test. Both Wald and LR tests for the geographic weight matrix rejected the degeneration hypothesis into the SEM, and the economic, geographic nested matrix passed all significance tests. Therefore, the SDM and SAR are considered for analysis in this study.

Table 8

Table 8. Results of spatial econometrics tests.

5.5.3 Spatial econometrics regression

Table 9 presents the results of the spatial spillover effects across two weight matrices and various spatial models. The estimated coefficient of AI's spatial effect on GEE is significantly positive, indicating that AI in a given region enhances the GEE of neighboring areas. This suggests that AI not only boosts local GEE but also exerts a spatial spillover effect, radiating and driving improvements in the GEE of neighboring regions.

Table 9

Table 9. Spatial econometric analysis results.

6 Non-linear impact of artificial intelligence on green economic efficiency

6.1 The non-linear impact of artificial intelligence on green economic efficiency under policy governance

The results in Table 10 show that different threshold effects of various environmental regulation policies influence AI-empowered GEE. MER and CER satisfy the single threshold significance test, while IER meets the single and double threshold significance test criteria.

Table 10

Table 10. Testing the threshold effects of heterogeneous environmental regulations.

This paper further examines the threshold effects of heterogeneous environmental regulation, as presented in Table 11. When MER exceeds the threshold, the regression coefficient decreases from 0.441 to 0.238, weakening the positive impact of AI on GEE. This finding supports H2a. Conversely, when the intensity of CER surpasses the first threshold, the coefficient rises from 0.101 to 0.249, enhancing AI's positive impact on GEE, which supports H2b. Similarly, for IER, when its intensity exceeds the first threshold but remains below the second, the coefficient increases from 0.123 to 0.269. Upon crossing the second threshold, the coefficient increases to 0.408, which is significant at the 1% level. These results indicate that a higher intensity of IER strengthens AI's effect on GEE, supporting H2c.

Table 11

Table 11. Threshold effect regression results of heterogeneous environmental regulations.

6.2 The non-linear impact of artificial intelligence on green economic efficiency under technological governance

Table 12 results show that GEE enabled by AI is influenced by different threshold effects of heterogeneous green technology innovations. SUG has passed single and double threshold significance tests, while SYG has passed the single threshold significance test.

Table 12

Table 12. Testing the threshold effects of heterogeneous green technology innovations.

This study investigates the thresholds of heterogeneous green technologies innovation, as summarized in Table 13. When SUG's intensity exceeds the first threshold but remains below the second, the coefficient decreases from 0.555 to 0.122. Beyond the second threshold, the coefficient rises to 0.319. These results indicate that the marginal effect of AI on GEE diminishes significantly after SUG surpasses the first threshold but partially recovers upon exceeding the second threshold. However, the effect remains weaker than observed before crossing the first threshold, supporting H3a. For SYG, when their intensity surpasses the first threshold, the coefficient increases from 0.084 to 0.249. This finding suggests that SYG enhances the marginal effect of AI on GEE, supporting H3b. Additionally, the threshold effects of SUG (0.433 and 0.197) are greater than those of SYG (0.165), supporting H3c.

Table 13

Table 13. Threshold regression results of heterogeneous green technology innovations.

6.3 The non-linear impact of artificial intelligence on green economic efficiency under legal governance

The results in Table 14 illustrate that AIP and JIP get through the single threshold significance test, signaling that the efficiency of AI-empowered GEE is affected by the threshold effect of AIP and JIP.

Table 14

Table 14. Testing the threshold effects of heterogeneous intellectual property protection.

This paper further examines the role of heterogeneous intellectual property protection thresholds, as depicted in Table 15. When the intensity of AIP exceeds the first threshold, the coefficient decreases from 0.423 to 0.165, significantly diminishing the marginal effect of AI on GEE. This finding supports H4a. Similarly, when the JIP surpasses the first threshold, the coefficient declines from 0.232 to 0.150, reducing AI's marginal effect on GEE and supporting H4b. Moreover, the threshold effect of AIP (0.258) is greater than that of JIP (0.082), providing evidence for H4c.

Table 15

Table 15. Threshold effect regression results of heterogeneous intellectual property protection.

7 Discussion and conclusion

7.1 Discussion

The significant contribution of AI to enhancing green economic efficiency aligns with the findings of Luo and Feng (2024) and Wang et al. (2024a). However, it contrasts with the conclusions of Lee et al. (2024), who consider the initial application stage of AI. During this stage, AI may negatively affect the energy transition. Over time, as AI progresses beyond its early implementation phase, its positive impacts are expected to outweigh the adverse effects associated with technological development and scale, ultimately driving the energy transition. The conclusion that variations in environmental regulation influence the marginal impacts of AI-empowered GEE aligns with the findings of Sun et al. (2023) and Wang et al. (2023b). However, it may diverge from the conclusions of You et al. (2024), who overlooked the cost impacts associated with environmental regulation. The conclusion regarding the marginal effect of heterogeneous green technology innovation on AI-empowered GEE aligns with the findings of Lian et al. (2022) and Chen et al. (2024a). However, Wang et al. (2024b) warn that green innovation may escalate corporate debt default risks, thereby affecting financial performance and reducing GEE. Excessive AIP and JIP may reduce the marginal effects of AI on GEE. Wang et al. (2024c) and Hudson and Minea (2013) have highlighted the non-linear impact of intellectual property protection on the green innovation capabilities of the manufacturing sector and innovation at large. Hu et al. (2023) particularly underscore the importance of stringent regulatory enforcement of intellectual property protection levels to foster innovation. Various scholars have examined the relationship between AI and GEE across different geographical contexts. Some have demonstrated that AI may enhance GEE in economies with development levels similar to China's. For instance, Akram et al. (2024) found that AI supports high-quality economic development in emerging economies by facilitating energy transition and green technology innovation. Similarly, Salman et al. (2024) confirmed that technological progress has raised the carbon neutrality rates of G20 countries. On a global scale, other studies provide evidence that AI can enhance GEE. Tao (2024) reported that AI has boosted global green productivity. Additionally, Lee et al. (2024) confirmed that AI has promoted energy transition and reduced carbon emissions across 69 countries. Despite variations in countries and economic conditions, AI promotes green and sustainable development through technological advancements, thereby extending the conclusions of this paper.

7.2 Conclusion and practical implications

7.2.1 Conclusion

This paper reveals the following findings: (1) AI significantly improves GEE, substantially impacting eastern, more open, and technologically advanced provinces. (2) In policy governance, excessive MER reduces AI's marginal effect, whereas strengthening CER and IER increases AI's marginal effect. (3) In technological governance, SUG reduces AI's marginal effect, while SYG may increase it, though SUG has a greater threshold effect than symbolic innovation. (4) In legal governance, AIP and JIP may reduce AI's marginal effect, with AIP showing a larger threshold effect than judicial protection. This study integrates AI with GEE within a comprehensive governance framework, providing a novel perspective on the heterogeneous and threshold effects across various governance types.

7.2.2 Practical implications

This paper draws the following conclusions based on previous research findings: First, the government should enhance policy support for technological innovation in AI. Second, the government should employ a combination of governance measures to maximize AI's empowering effects. In the policy governance process, the government should carefully leverage the roles of CER, fully utilize IER, and control MER intensity. In technical governance, enterprises should prioritize SUG over SYG while maintaining moderate investment in green technology R&D. For legal governance; the government should continuously strengthen intellectual property protection, ensuring that AIP is balanced to guide behavior without excessive intervention. Third, the green economic efficiency of spatially connected regions is influenced by the radiating and driving effects of local AI development; therefore, regions should facilitate the flow of human, technical, knowledge, and technological resources to support regional green synergistic development.

7.3 Limitations and prospects

This paper has shortcomings in the following aspects, which deserve to be supplemented and improved in subsequent studies. First, the research is limited to China, and future studies might consider expanding this geographical scope. Second, upon collecting all variable data, numerous missing values were identified after 2020. Additionally, due to the COVID-19 outbreak, substantial changes in policy and market conditions may have rendered data from the pandemic period incomparable with data from other periods. Consequently, the data has only been updated through 2020 to ensure availability and reliability. Third, environmental protection outputs are not included in the indicator system when assessing GEE. Future research should further investigate the direct impacts of AI on green economic efficiency, specifically through the lenses of industrial structure and green innovation. Fourth, this study primarily focuses on the efficiency of AI during the operational phase without comprehensively evaluating the entire lifecycle of AI models, including the resource-intensive training and deployment phases that may increase carbon emissions. Future research should assess the green economic efficiency across the whole lifecycle of AI models.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: International Federation of Robotics.

Author contributions

ZS: Data curation, Formal analysis, Funding acquisition, Methodology, Resources, Software, Writing – original draft, Writing – review & editing. YD: Investigation, Methodology, Resources, Writing – original draft.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This work was supported by the National Social Science Foundation of China (Project No. 23BGL063) and The Chongqing Municipal Education Commission Humanities and Social Sciences Research Planning Project (Project No: 24SKGH419).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Abulibdeh, A., Zaidan, E., and Abulibdeh, R. (2024). Navigating the confluence of artificial intelligence and education for sustainable development in the era of industry 4.0: challenges, opportunities, and ethical dimensions. J. Clean. Product. 437:140527. doi: 10.1016/j.jclepro.2023.140527

Crossref Full Text | Google Scholar

Adnan, M., Xiao, B., Ali, M. U., Bibi, S., Yu, H., Xiao, P., et al. (2024). Human inventions and its environmental challenges, especially artificial intelligence: new challenges require new thinking. Environ. Chall. 16:100976. doi: 10.1016/j.envc.2024.100976

Crossref Full Text | Google Scholar

Ahmed, I., Jeon, G., and Piccialli, F. (2022). From artificial intelligence to explainable artificial intelligence in industry 4.0: a survey on what, how, and where. Ieee Transact. Indus. Inform. 18, 5031–5042. doi: 10.1109/TII.2022.3146552

Crossref Full Text | Google Scholar

Akram, R., Li, Q., Srivastava, M., Zheng, Y., and Ifran, M. (2024). Nexus between green technology innovation and climate policy uncertainty: unleashing the role of artificial intelligence in an emerging economy. Technol. Forecast. Soc. Change 209:123820. doi: 10.1016/j.techfore.2024.123820

Crossref Full Text | Google Scholar

Akter, S., Babu, M. M., Hani, U., Sultana, S., Bandara, R., Grant, D., et al. (2024). Unleashing the power of artificial intelligence for climate action in industrial markets. Indust. Market. Manage. 117, 92–113. doi: 10.1016/j.indmarman.2023.12.011

Crossref Full Text | Google Scholar

Amsler, L. B. (2016). Collaborative governance: integrating management, politics, and law. Publ. Admin. Rev. 76, 700–711. doi: 10.1111/puar.12605

Crossref Full Text | Google Scholar

Aparicio, J., Ortiz, L., and Pastor, J. T. (2017). Measuring and decomposing profit inefficiency through the Slacks-Based measure. Eur. J. Oper. Res. 260, 650–654. doi: 10.1016/j.ejor.2016.12.038

Crossref Full Text | Google Scholar

Bibri, S. E., Alexandre, A., Sharifi, A., and Krogstie, J. (2023). Environmentally sustainable smart cities and their converging AI, IoT, and big data technologies and solutions: an integrated approach to an extensive literature review. Energy Inform. 6:9. doi: 10.1186/s42162-023-00259-2

PubMed Abstract | Crossref Full Text | Google Scholar

Bibri, S. E., Krogstie, J., Kaboli, A., and Alahi, A. (2024). Smarter eco-cities and their leading-edge artificial intelligence of things solutions for environmental sustainability: a comprehensive systematic review. Environ. Sci. Ecotechnol. 19:100330. doi: 10.1016/j.ese.2023.100330

PubMed Abstract | Crossref Full Text | Google Scholar

Cai, D., Aziz, G., Sarwar, S., Alsaggaf, M. I., and Sinha, A. (2024). Applicability of denoising-based artificial intelligence to forecast the environmental externalities. Geosci. Front. 15:101740. doi: 10.1016/j.gsf.2023.101740

Crossref Full Text | Google Scholar

Chen, C., Hu, Y., Karuppiah, M. M. K., and Kumar, P. M. (2021). Artificial intelligence on economic evaluation of energy efficiency and renewable energy technologies. Sustain. Energy Technol. Assess. 47:101358. doi: 10.1016/j.seta.2021.101358

Crossref Full Text | Google Scholar

Chen, C., Li, W.-B., and Zhang, H. (2024a). How do property rights affect corporate ESG performance? The moderating effect of green innovation efficiency. Finance Res. Lett. 64:105476. doi: 10.1016/j.frl.2024.105476

Crossref Full Text | Google Scholar

Chen, C., Pinar, M., and Roman-Collado, R. (2024b). Green innovation and energy efficiency: moderating effect of institutional quality based on the threshold model. Environ. Resour. Econ. 87, 3063–3094. doi: 10.1007/s10640-024-00902-w

Crossref Full Text | Google Scholar

Chen, S., and Yang, Q. (2024). Renewable energy technology innovation and urban green economy efficiency. J. Environ. Manage. 353:120130. doi: 10.1016/j.jenvman.2024.120130

PubMed Abstract | Crossref Full Text | Google Scholar

Dao, N. B., Dogan, B., Ghosh, S., Kazemzadeh, E., and Radulescu, M. (2024). Toward sustainable ecology: how do environmental technologies, green financial policies, energy uncertainties, and natural resources rents matter? Clean Technol. Environ. Policy. doi: 10.1007/s10098-024-02887-y

Crossref Full Text | Google Scholar

DiMaggio, P. J., and Powell, W. W. (1983). The iron cage revisited: institutional isomorphism and collective rationality in organizational fields. Am. Sociol. Rev. 48, 147–160. doi: 10.2307/2095101

Crossref Full Text | Google Scholar

Fang, G., Gao, Z., Wang, L., and Tian, L. (2022). How does green innovation drive urban carbon emission efficiency?—Evidence from the Yangtze River Economic Belt. J. Clean. Product. 375:134196. doi: 10.1016/j.jclepro.2022.134196

Crossref Full Text | Google Scholar

Feng, C., Ye, X., Li, J., and Yang, J. (2024). How does artificial intelligence affect the transformation of China's green economic growth? An analysis from internal-structure perspective. J. Environ. Manage. 351:119923. doi: 10.1016/j.jenvman.2023.119923

PubMed Abstract | Crossref Full Text | Google Scholar

Gao, Z., Li, L., and Hao, Y. (2024). Resource industry dependence and high-quality economic development of Chinese style: reexamining the effect of the “Resource Curse”. Struct. Change Econ. Dyn. 68, 1–16. doi: 10.1016/j.strueco.2023.09.013

Crossref Full Text | Google Scholar

Hall, B. H., and Helmers, C. (2013). Innovation and diffusion of clean/green technology: can patent commons help? J. Environ. Econ. Manage. 66, 33–51. doi: 10.1016/j.jeem.2012.12.008

Crossref Full Text | Google Scholar

Hu, X., Zhang, Z., and Lv, C. (2023). The impact of technological transformation on basic research results: the moderating effect of intellectual property protection. J. Innov. Knowl. 8:100443. doi: 10.1016/j.jik.2023.100443

Crossref Full Text | Google Scholar

Huang, L., Zhang, H., Si, H., and Wang, H. (2023). Can the digital economy promote urban green economic efficiency? Evidence from 273 cities in China. Ecol. Indicat. 155:110977. doi: 10.1016/j.ecolind.2023.110977

Crossref Full Text | Google Scholar

Hudson, J., and Minea, A. (2013). Innovation, intellectual property rights, and economic development: a unified empirical investigation. World Dev. 46, 66–78. doi: 10.1016/j.worlddev.2013.01.023

Crossref Full Text | Google Scholar

Kazemzadeh, E., Fuinhas, J. A., Shirazi, M., Koengkan, M., and Silva, N. (2023). Does economic complexity increase energy intensity? Energy Effic. 16:29. doi: 10.1007/s12053-023-10104-w

Crossref Full Text | Google Scholar

Koengkan, M., Fuinhas, J. A., and Kazemzadehc, E. (2024). Do financial incentive policies for renewable energy development increase the economic growth in Latin American and Caribbean countries? J. Sustain. Finance Invest. 14, 161–183. doi: 10.1080/20430795.2022.2031849

Crossref Full Text | Google Scholar

Konya, A., and Nematzadeh, P. (2024). Recent applications of AI to environmental disciplines: a review. Sci. Total Environ. 906:167705. doi: 10.1016/j.scitotenv.2023.167705

PubMed Abstract | Crossref Full Text | Google Scholar

Lee, C.-C., Yan, J., and Li, T. (2024). Ecological resilience of city clusters in the middle reaches of Yangtze river. J. Clean. Product. 443:141082. doi: 10.1016/j.jclepro.2024.141082

Crossref Full Text | Google Scholar

Li, H., and Shi, J.-F. (2014). Energy efficiency analysis on Chinese industrial sectors: an improved Super-SBM model with undesirable outputs. J. Clean. Product. 65, 97–107. doi: 10.1016/j.jclepro.2013.09.035

Crossref Full Text | Google Scholar

Li, H., and Xu, R. (2023). How does digital finance affect the efficiency of urban green economies? Evidence from China. Finance Res. Lett. 58:104595. doi: 10.1016/j.frl.2023.104595

Crossref Full Text | Google Scholar

Li, Q., Zhang, J., Feng, Y., Sun, R., and Hu, J. (2024). Towards a high-energy efficiency world: assessing the impact of artificial intelligence on urban energy efficiency. J. Clean. Product. 461:142593. doi: 10.1016/j.jclepro.2024.142593

Crossref Full Text | Google Scholar

Li, W., and Yuan, B. (2024). Intellectual property protection and M&A performance. Finance Res/ Lett. 64:105395. doi: 10.1016/j.frl.2024.105395

PubMed Abstract | Crossref Full Text | Google Scholar

Lian, G., Xu, A., and Zhu, Y. (2022). Substantive green innovation or symbolic green innovation? The impact of ER on enterprise green innovation based on the dual moderating effects. J. Innovat. Knowl. 7:100203. doi: 10.1016/j.jik.2022.100203

Crossref Full Text | Google Scholar

Lin, B., and Zhou, Y. (2022). Measuring the green economic growth in China: influencing factors and policy perspectives. Energy 241:122518. doi: 10.1016/j.energy.2021.122518

Crossref Full Text | Google Scholar

Liu, J., Chang, H., Forrest, J. Y. L., and Yang, B. (2020). Influence of artificial intelligence on technological innovation: evidence from the panel data of china's manufacturing sectors. Technol. Forecast. Soc. Change 158:120142. doi: 10.1016/j.techfore.2020.120142

Crossref Full Text | Google Scholar

Liu, X., and Zhang, X. (2021). Industrial agglomeration, technological innovation and carbon productivity: evidence from China. Resour. Conserv. Recycl. 166:105330. doi: 10.1016/j.resconrec.2020.105330

Crossref Full Text | Google Scholar

Liu, Y., and Dong, F. (2021). How technological innovation impacts urban green economy efficiency in emerging economies: a case study of 278 Chinese cities. Resour. Conserv. Recycl. 169:105534. doi: 10.1016/j.resconrec.2021.105534

Crossref Full Text | Google Scholar

Luo, Q., and Feng, P. (2024). Exploring artificial intelligence and urban pollution emissions: “speed bump” or “accelerator” for sustainable development? J. Clean. Product. 463, 142739. doi: 10.1016/j.jclepro.2024.142739

Crossref Full Text | Google Scholar

Lv, L., and Chen, Y. (2024). The collision of digital and green: digital transformation and green economic efficiency. J. Environ. Manage. 351:119906. doi: 10.1016/j.jenvman.2023.119906

PubMed Abstract | Crossref Full Text | Google Scholar

Naeem, R., Kohtamaeki, M., and Parida, V. (2024). Artificial intelligence enabled product-service innovation: past achievements and future directions. Rev. Manage. Sci. doi: 10.1007/s11846-024-00757-x

Crossref Full Text | Google Scholar

Nahar, S. (2024). Modeling the effects of artificial intelligence (AI)-based innovation on sustainable development goals (SDGs): applying a system dynamics perspective in a cross-country setting. Technol. Forecast. Soc. Change 201:123203. doi: 10.1016/j.techfore.2023.123203

Crossref Full Text | Google Scholar

Olawade, D. B., Wada, O. Z., David-Olawade, A. C., Fapohunda, O., and Ling, J. (2024). Artificial intelligence potential for net zero sustainability: current evidence and prospects. Next Sustain. 4:100041. doi: 10.1016/j.nxsust.2024.100041

Crossref Full Text | Google Scholar

Porter, M. E., and van der Linde, C. (1995). Toward a new conception of the environment-competitiveness relationship. J. Econ. Perspect. 9, 97–118. doi: 10.1257/jep.9.4.97

Crossref Full Text | Google Scholar

Rasheed, M. Q., Yuhuan, Z., Haseeb, A., Ahmed, Z., and Saud, S. (2024a). Assessing green energy production and industrial excellence in Asian emerging economies in the context of industrial transformation and sustainable development. Environ. Dev. Sustain. doi: 10.1007/s10668-024-05099-y

Crossref Full Text | Google Scholar

Rasheed, M. Q., Yuhuan, Z., Haseeb, A., Ahmed, Z., and Saud, S. (2024b). Asymmetric relationship between competitive industrial performance, renewable energy, industrialization, and carbon footprint: does artificial intelligence matter for environmental sustainability? Appl. Energy 367:123346. doi: 10.1016/j.apenergy.2024.123346

Crossref Full Text | Google Scholar

Rasheed, M. Q., Yuhuan, Z., Haseeb, A., Ahmed, Z., and Saud, S. (2024c). Information communication technology, economic growth, natural resources, and renewable energy production: evaluating the asymmetric and symmetric impacts of artificial intelligence in robotics and innovative economies. J. Clean. Product. 447:141466. doi: 10.1016/j.jclepro.2024.141466

Crossref Full Text | Google Scholar

Salman, M., Wang, G., Qin, L., and He, X. (2024). G20 roadmap for carbon neutrality: the role of Paris agreement, artificial intelligence, and energy transition in changing geopolitical landscape. J. Environ. Manage. 367:122080. doi: 10.1016/j.jenvman.2024.122080

PubMed Abstract | Crossref Full Text | Google Scholar

Shaamala, A., Yigitcanlar, T., Nili, A., and Nyandega, D. (2024). Algorithmic green infrastructure optimisation: review of artificial intelligence driven approaches for tackling climate change. Sustain. Cities Soc. 101:105182. doi: 10.1016/j.scs.2024.105182

Crossref Full Text | Google Scholar

Shen, Q., Pan, Y., and Feng, Y. (2023). Identifying and assessing the multiple effects of informal environmental regulation on carbon emissions in China. Environ. Res. 237:116931. doi: 10.1016/j.envres.2023.116931

PubMed Abstract | Crossref Full Text | Google Scholar

Shuai, S., and Fan, Z. (2020). Modeling the role of environmental regulations in regional green economy efficiency of China: empirical evidence from super efficiency DEA-Tobit model. J. Environ. Manage. 261:110227. doi: 10.1016/j.jenvman.2020.110227

PubMed Abstract | Crossref Full Text | Google Scholar

Song, H., and Chen, W. (2023). Can two-way foreign direct investment promote green innovation capability in manufacturing? The threshold role of intellectual property protection. J. Clean. Product. 425:139035. doi: 10.1016/j.jclepro.2023.139035

Crossref Full Text | Google Scholar

Song, H. Y., Hou, J., and Zhang, Y. (2022). Patent protection: does it promote or inhibit the patented technological knowledge diffusion? Scientometrics 127, 2351–2379. doi: 10.1007/s11192-022-04348-z

Crossref Full Text | Google Scholar

Song, W., Han, X., and Liu, Q. (2024). Patterns of environmental regulation and green innovation in China. Struct. Change Econ. Dyn. 71, 176–192. doi: 10.1016/j.strueco.2024.07.006

Crossref Full Text | Google Scholar

Sun, J., Zhai, N., Miao, J., Mu, H., and Li, W. (2023). How do heterogeneous environmental regulations affect the sustainable development of marine green economy? Empirical evidence from China's coastal areas. Ocean Coastal Manage. 232:106448. doi: 10.1016/j.ocecoaman.2022.106448

Crossref Full Text | Google Scholar

Tao, M. (2024). Digital brains, green gains: artificial intelligence's path to sustainable transformation. J. Environ. Manage. 370:122679. doi: 10.1016/j.jenvman.2024.122679

PubMed Abstract | Crossref Full Text | Google Scholar

Tian, H., Zhao, L., Li, Y., and Wang, W. (2023). Can enterprise green technology innovation performance achieve “corner overtaking” by using artificial intelligence?—Evidence from Chinese manufacturing enterprises. Technol. Forecast. Soc. Change 194:122732. doi: 10.1016/j.techfore.2023.122732

Crossref Full Text | Google Scholar

Tone, K. (2001). A slacks-based measure of efficiency in data envelopment analysis. Eur. J. Operat. Res. 130, 498–509. doi: 10.1016/S0377-2217(99)00407-5

Crossref Full Text | Google Scholar

Wang, J., and Zhou, X. (2024). Measurement and synergistic evolution analysis of economic resilience and green economic efficiency: evidence from five major urban agglomerations, China. Appl. Geograp. 168:103302. doi: 10.1016/j.apgeog.2024.103302

Crossref Full Text | Google Scholar

Wang, L., Long, Y., and Li, C. (2022). Research on the impact mechanism of heterogeneous environmental regulation on enterprise green technology innovation. J.Environ. Manage. 322:116127. doi: 10.1016/j.jenvman.2022.116127

PubMed Abstract | Crossref Full Text | Google Scholar

Wang, Q., Zhang, F., Li, R., and Sun, J. (2024a). Does artificial intelligence promote energy transition and curb carbon emissions? The role of trade openness. J. Clean. Product. 447:141298. doi: 10.1016/j.jclepro.2024.141298

Crossref Full Text | Google Scholar

Wang, X., Chu, B., Ding, H., and Chiu, A. S. F. (2023a). Impacts of heterogeneous environmental regulation on green transformation of China's iron and steel industry: evidence from dynamic panel threshold regression. J. Clean. Product. 382:135214. doi: 10.1016/j.jclepro.2022.135214

Crossref Full Text | Google Scholar

Wang, X., Guo, Y., and Fu, S. (2024b). Will green innovation strategies trigger debt default risk? Evidence from listed companies in China. Finance Res. Lett. 62:105216. doi: 10.1016/j.frl.2024.105216

Crossref Full Text | Google Scholar

Wang, Y., Shi, J., and Qu, G. (2024c). Research on collaborative innovation cooperation strategies of manufacturing digital ecosystem from the perspective of multiple stakeholders. Comp. Indus. Eng. 190:110003. doi: 10.1016/j.cie.2024.110003

Crossref Full Text | Google Scholar

Wang, Z.-R., Fu, H.-Q., and Ren, X.-H. (2023b). The impact of political connections on firm pollution: new evidence based on heterogeneous environmental regulation. Petrol. Sci. 20, 636–647. doi: 10.1016/j.petsci.2022.10.019

Crossref Full Text | Google Scholar

Xu, B. (2024). Environmental regulations, technological innovation, and low carbon transformation: a case of the logistics industry in China. J. Clean. Product. 439:140710. doi: 10.1016/j.jclepro.2024.140710

Crossref Full Text | Google Scholar

Yang, H., Tang, M., Chao, X., and Li, P. (2023). How environmental regulation influences the green economy efficiency of resource-based cities: an empirical study from China. Environ. Dev. Sustain. 27249–27276. doi: 10.1007/s10668-023-03758-0

PubMed Abstract | Crossref Full Text | Google Scholar

Yang, H., Xu, X., and Zhang, F. (2022). Industrial co-agglomeration, green technological innovation, and total factor energy efficiency. Environ. Sci. Pollut. Res. 29, 62475–62494. doi: 10.1007/s11356-022-20078-4

PubMed Abstract | Crossref Full Text | Google Scholar

Yang, T., Yi, X., Lu, S., Johansson, K. H., and Chai, T. (2021a). Intelligent manufacturing for the process industry driven by industrial artificial intelligence. Engineering 7, 1224–1230. doi: 10.1016/j.eng.2021.04.023

Crossref Full Text | Google Scholar

Yang, X., Wu, H., Ren, S., Ran, Q., and Zhang, J. (2021b). Does the development of the internet contribute to air pollution control in China? Mechanism discussion and empirical test. Struct. Change Econ. Dyn. 56, 207–224. doi: 10.1016/j.strueco.2020.12.001

Crossref Full Text | Google Scholar

You, Z., Hou, G., and Wang, M. (2024). Heterogeneous relations among environmental regulation, technological innovation, and environmental pollution. Heliyon 10:e28196. doi: 10.1016/j.heliyon.2024.e28196

PubMed Abstract | Crossref Full Text | Google Scholar

Zhang, H., Tao, L., Yang, B., and Bian, W. (2023). The relationship between public participation in environmental governance and corporations? Environmental violations. Finance Res. Lett. 53:103676. doi: 10.1016/j.frl.2023.103676

Crossref Full Text | Google Scholar

Zhang, H., Xu, T., and Feng, C. (2022). Does public participation promote environmental efficiency? Evidence from a quasi-natural experiment of environmental information disclosure in China. Energy Econ. 108:105871. doi: 10.1016/j.eneco.2022.105871

Crossref Full Text | Google Scholar

Zhang, Z., Hou, Y. Z., and Li, M. (2024a). From symbolic to substantive green innovation: how does ESG ratings optimize corporate green innovation structure. Finance Res. Lett. 63:105401. doi: 10.1016/j.frl.2024.105401

Crossref Full Text | Google Scholar

Zhang, Z., Li, R., Song, Y., and Sahut, J.-M. (2024b). The impact of environmental regulation on the optimization of industrial structure in energy-based cities. Res. Int. Business Finance 68:102154. doi: 10.1016/j.ribaf.2023.102154

Crossref Full Text | Google Scholar

Zhao, P., Gao, Y., and Sun, X. (2022). How does artificial intelligence affect green economic growth?—Evidence from China. Sci. Total Environ. 834:155306. doi: 10.1016/j.scitotenv.2022.155306

PubMed Abstract | Crossref Full Text | Google Scholar

Zhou, C., Chenyi Shi, S. W., and Zhang, G. (2018). Estimation of eco-efficiency and its influencing factors in Guangdong province based on Super-SBM and panel regression models. Ecol. Indicat. 86, 67–80. doi: 10.1016/j.ecolind.2017.12.011

Crossref Full Text | Google Scholar

Zhou, W., Zhang, Y., and Li, X. (2024a). Artificial intelligence, green technological progress, energy conservation, and carbon emission reduction in China: an examination based on dynamic spatial Durbin modeling. J. Clean. Product. 446:141142. doi: 10.1016/j.jclepro.2024.141142

Crossref Full Text | Google Scholar

Zhou, W., Zhuang, Y., and Chen, Y. (2024b). How does artificial intelligence affect pollutant emissions by improving energy efficiency and developing green technology. Energy Econ. 131:107355. doi: 10.1016/j.eneco.2024.107355

Crossref Full Text | Google Scholar