The production of enterprises consumes a large amount of resources and energy, and discharges a lot of waste gas, wastewater and so on that cause serious damage and pollution to the environment. GTFP has attracted widespread attention, and many scholars have researched GTFP. Based on the meta-frontier theory and data envelopment analysis method, C Fang et al. calculated the green total factor productivity of extractive industries in China from 2006 to 2017, and analyzed its influencing factors [13]. Based on the comprehensive control method (SCM) using microdata, Jiang Yufan et al. studied the net impact of the establishment of China’s (Shanghai) pilot free trade zone on Shanghai’s green total factor productivity [14]. The study measured green total factor productivity (GTFP) to discuss whether factor market distortions inhibit the green development of China’s economy. Boqiang Lin et al. found that factor market distortion had a negative impact on China’s GTFP growth [15]. Wu used an SFA basic model to analyze the temporal and spatial dynamic evolution trend of marine economic green production efficiency in coastal areas of China [16]. D Liu et al. used the Super-SBM model to calculate China’s agricultural green total factor productivity. On this basis, the kernel density estimation method was used to study its dynamic evolution trend, and the panel data model was used to conduct empirical research on the influencing factors of China’s agricultural green total factor productivity [17]. Erdmann et al. studied the impact of ICT on greenhouse gas emissions through a scenario analysis, and found that in most scenarios, ICT reduced greenhouse gas emissions [18].

There are few studies on the impact of intelligence on green total factor productivity, and some have studied the impact of intelligent on production efficiency. Scholars’ research on the relationship between intelligence and TFP is mainly divided into two categories.

On the one hand, some scholars thought that intelligence can promote the growth of manufacturing TFP. Acemoglu and Restrepo studied the impact of automation on labor productivity and employment based on the theory of endogenous growth, and found that automation reduced the demand for labor through the use of cheap capital, and productivity was greatly improved [19]. Kromann et al. used panel data from 10 manufacturing industries in nine countries from 2004 to 2007 to study the impact of increased automation on productivity. The results showed that large-scale use of industrial robots had a significant effect on labor productivity [20]. Brynjolfsson et al. used survey data from 179 listed companies to analyze the impact of business practices and information-technology investment on production efficiency. The study showed that the output and productivity of companies that used data-driven decision-making were higher than for those that used other investment and information technologies by 5–6% [21]. Graetz and Michaels used panel data to empirically study the impact of automation on the economy. The research showed that the use of industrial robots significantly increased labor productivity, increasing GDP and labor productivity by 0.37% and 0.36%, respectively [22]. Feng Zhaokui researched the workshop of the Japanese automobile manufacturing industry and found that with the increasingly low price of robots, the ratio of robots replacing workers in production lines was increasing, which greatly reduced the labor required for daily production and improves labor productivity [23]. Commander et al. conducted research on companies in Brazil and India and found that information and communication technology had a positive effect on the productivity in both countries, and this effect was highly adaptable to different testing methods [24]. He Zhao et al. studied the application of AR/VR technologies in in the field of managing human resources (HRs) and basic operational corporate business processes. They found that AR/VR technologies could optimize business operation processes [25].

On the other hand, some scholars believe that intelligence has not significantly promoted total factor productivity development, or even inhibited such development. Cowen pointed out that the development of the Internet and computer technology does not promote productivity significantly. He believed that in the general perception, technological progress can effectively promote the increase of productivity. But when it comes to computer technology, this experience often turns out to be wrong [26]. Brynjolfsson et al. believed that there are four possible explanations for the artificial intelligence productivity paradox: false expectations, measurement errors, centralized distribution and rent dissipation, and implementation and restructuring delays [27]. Acemoglu and Restrepo pointed out that excessive automation is the main reason for suppressing productivity growth. Excessive automation not only directly leads to low production efficiency, but also may indirectly inhibit productivity growth through waste of resources and labor mismatch [28]. Acemoglu et al. made the research on the impact of the use of industrial robots in 19 industries on the U.S. labor market, and found that the large-scale application of industrial robots has a significant negative impact on employment and wages [29]. Stiroh conducted research based on the US manufacturing industry and believed that there was no evidence that the spillover effects of information technology would bring about total factor productivity growth [30].

Current literature contains rich research on intelligence and GTFP. However, there is a lack of direct research on the relationship between intelligence and the quality of economic development, and there are no studies on the relationship between intelligence and economic green development. Therefore, this article comprehensively considers the three factors of intelligence, economic development, and the environment, analyzes the internal mechanism of intelligence on the green development of the manufacturing industry, and provides a new perspective for the green transformation and upgrading of China’s economy. At the same time, this article studies the internal connection between the integrated development of “intelligence + manufacturing” and the transformation and upgrading of the manufacturing industry, which has important practical significance for improving the intelligence level of China’s manufacturing industry and promoting the green manufacturing industry.

Regarding the development of GTFP and intelligence in China, Figure 1 shows the actual situation of GTFP and intelligence in China from 2008 to 2017.The specific calculation method will be introduced in detail below. As can be seen, GTFP and intelligence had similar trends from 2008 to 2017. At present, there are few studies of the relationship between intelligence and green total factor productivity in manufacturing at home and abroad, but studies related to Internet technology, information technology, and environmental protection found that the Internet and informatization are beneficial in promoting environmental protection. The Internet can improve environmental quality by reducing energy consumption and reducing pollution emissions. The use of the Internet is conducive to the reduction of energy intensity and the improvement of energy efficiency in the production sector, thereby reducing energy consumption [31,32]. The development of information and communication technology can suppress the level of carbon dioxide emissions, and the use of Internet platforms for transactions can reduce carbon dioxide emissions in developed countries [33,34]. We can make the hypothesis that intelligence has a positive relationship with GTFP.

where α and β are the contribution coefficients in intelligence and GTFP. We determined that α and β are both equal to 0.5, considering that they are both equally important in manufacturing, and interact equally within the system. G is the coupling degree of these two subsystems, V is the comprehensive index of these subsystems, and U is the coupling coordination degree of the system. According to the related research, U can be divided into four stages: 𝑈∈(0,0.3)𝑈∈(0,0.3), which indicates the low coordination stage; 𝑈∈(0.3,0.5)𝑈∈(0.3,0.5), which indicates the moderate coordination stage; 𝑈∈(0.5,0.8)𝑈∈(0.5,0.8), which indicates the high coordination stage; and 𝑈∈(0.8,0.1)𝑈∈(0.8,0.1), which indicates the extreme coordination stage.

The degree of coupling coordination between the level of intelligence and the GTFP of the manufacturing industry is shown in Figure 2. In 2008, the coupling coordination degree of intelligent and GTFP was 0.501, which was in the middle coordination stage. In 2017, the coupling coordination degree of intelligence and GTFP was 0.726, which was in a highly coordinated stage. The degree of coupling and coordination from 2008 to 2017 exceeded 0.5, indicating that there was a close relationship between intelligence and the GTFP of manufacturing.

Considering the impact of intelligence on GTFP, it is essential to consider both the input and output aspects. Including capital, traditional labor production factors input and energy input. Output includes expected output GDP and undesired output environmental pollutant emissions. It is generally believed that the lower the input, the higher the expected output, and that lower undesired output means the GTFP increases faster. In addition, green total factor productivity can be divided into the forms of technical efficiency and technological progress [35,36,37]. Based on the Malmquist theory, this article analyzes the mechanism of the impact of intelligence on the GTFP of the manufacturing industry from the perspectives of technological progress and technical efficiency. This provides a realistic basis for our research on the relationship between intelligence and GTFP.

In order to comprehensively analyze the impact of intelligence on the GTFP of China’s manufacturing, and to explore the source of manufacturing GTFP changes, the empirical analysis of this paper was mainly divided into two parts: On the one hand, intelligence affects the overall regression of China’s manufacturing GTFP, and the regression of the GTFP decomposition item. On the other hand, this paper analyzes the influence of intelligence on the GTFP of manufacturing in China’s eastern, central, and western regions, and explores the regional heterogeneity of intelligence on the GTFP of China’s manufacturing industry. This study used Eviews10 to estimate the model; the results are shown in the Table 2.

According to Table 3, due to the different levels of economic and social development in the eastern, central, and western regions, the impact of intelligence on the green total factor productivity of manufacturing in different regions was also different. This conclusion was similar to that of the current papers [46,47,48].

It was found that intelligence had a positive and significant impact on the GTFP and technical efficiency of manufacturing in the eastern region, while the impact on the technical progress was not significant. Intelligence promoted the GTFP of manufacturing in the eastern region by improving the technical efficiency of manufacturing. This was mainly due to the fact that the manufacturing industry in eastern China mainly uses intelligent technology to improve the production efficiency of the manufacturing industry, instead of promoting technological upgrading through intelligent technology.

The impact of intelligence on the GTFP and technical efficiency of the manufacturing industry in the central region was positive and significant, and the impact on technological upgrading was not significant. This showed that the effect of intelligence on the GTFP of manufacturing in the central region was achieved by improving the technical efficiency of manufacturing enterprises in the central region. Compared with the eastern region, the development of manufacturing industry in the central region was underdeveloped. The application of intelligence in the manufacturing industry effectively promoted the improvement of the technical efficiency of manufacturing enterprises. Intelligence had no significant impact on the technological progress of the central region, which may be due to the low level of intelligence in this region.

Intelligence had no significant impact on the GTFP of manufacturing in the western region. This showed that intelligence has no obvious promotion effect on the GTFP of the manufacturing industry in the western region, the main reason for which was that the application of intelligence in the manufacturing industry in the west was still in its infancy at this stage, and the manufacturing industry in the west was still in the underdeveloped stage, and had a large gap between the manufacturing industries in the central and eastern regions.

Based on the analysis of the regression results of the whole country and the eastern and central regions, it was found that intelligence promoted the GTFP of manufacturing in the eastern and central regions by improving technical efficiency, and that intelligence had no significant effect on the improvement of technological progress. This is mainly because Chinese manufacturing companies use intelligent technology mainly at the application level, rather than for industrial upgrading and innovation.

5.3. Robustness Test

In order to consider the potential impact of the selection of variable indicators on the research results, and to ensure the reliability and stability of the research results, we measured intelligence from the perspective of output, and tested the robustness of the model. We used the total software business revenue to replace intelligence, while other variables remained unchanged.

According to a comparison of Table 2, Table 3 and Table 4, it was found that the variable signs and significance of the national overall and subregional regression results were basically unchanged, indicating that the national overall and subregional regression results were relatively stable; that is, intelligence promoted manufacturing GTFP through technical efficiency.

Table 4. Regression results of robustness tests.