人工智能技术对碳排放的影响

随着新一轮科技革命和产业变革的深入推进，人工智能技术在应对气候变化方面扮演重要角色，并赋能“碳达峰、碳中和”目标的实现。利用2006—2019年中国内地省级面板数据，考察人工智能技术对碳排放的影响。研究发现：人工智能技术与碳排放之间呈倒U型关系，即当人工智能技术达到一定阈值后，其碳减排效应逐渐凸显；人工智能技术主要通过提高能源利用效率实现碳减排；在东部和西部地区，人工智能技术与碳排放之间存在显著倒U型关系，而在中部地区，人工智能技术对碳排放发挥持续促进作用。因此，在推进“双碳”目标过程中，需要以绿色低碳发展为目标开发人工智能技术，正确引导人工智能在碳减排领域的研发应用，针对不同区域实际制定差异化发展政策。

The Effect of Artificial Intelligence Technology on Carbon Emissions

To actively address the issue of climate change, China has proposed a vision of "emission peak and carbon neutrality". Achieving "emission peak and carbon neutrality" is a comprehensive and profound economic and social systemic change that requires the joint efforts of the whole society. As a strategic technology for technological revolution and industrial change, Artificial Intelligence (AI) technologies will play an important role in addressing climate change and bring significant opportunities for low-carbon development. The process of achieving the "double carbon goal" is essentially a path of transformation to technology-intensive industry. However, the relationship between AI technologies and carbon emissions has been relatively under-discussed in academia compared to the much attention at the policy level and the flourishing development at the practical level. In particular, AI technologies may exhibit a more complex dual effect on carbon emissions. In view of this, this paper answers the question of whether AI technology can enable carbon neutrality and help reduce CO2 emissions. What are the mechanisms of AI technologies affecting carbon emissions? Are there any regional differences in the effect of AI technologies on carbon emissions? The purpose of this paper is to provide empirical evidence for the carbon reduction effect of AI technologies and to provide useful policy insights for achieving the goal of "double carbon".#br#In terms of theoretical research, this paper believes that the impact of AI technologies on carbon emissions may have a dual effect.On one hand, AI can increase total carbon emissions through direct energy consumption.At the same time, AI technology on carbon emissions may have a rebound effect.In addition, AI technologies has scale effect on economic growth, which will lead to the increase of total regional carbon emissions. On the other hand, AI technology can reduce carbon emissions by assisting decision-making, reshaping production and life styles, and facilitating low-carbon technological innovation.First, it can help reduce carbon emissions by assisting decision-making.Second, it can drive changes in production and consumption patterns to reduce carbon emissions.Third, AI reduces carbon emissions by enabling low-carbon technology innovation.To sum up, the relationship between AI technologies and carbon emissions is not a simple linear relationship, but depends on the combined effect of the above dual effects.In terms of influence mechanism, the effect of AI technologies on carbon emissions is mainly reflected in energy utilization efficiency.#br#This paper analyzes the effects, mechanisms and differences of AI technology on carbon emissions by manually collating AI patent data as a measure of AI technology, using panel data from 30 provinces in China from 2006 to 2019 as a research sample. The empirical results show that the effect of AI technology on carbon emissions shows an "inverted U-shaped" relationship at the national level as a whole. It means when the level of AI technology reaches a certain threshold, its carbon emission reduction effect gradually becomes prominent, and the robustness test results also reinforce this conclusion. The results of the intermediate effect test show that the effect of AI technology on carbon emissions is mainly achieved by affecting energy use efficiency. The heterogeneity analysis shows that there is a significant regional difference in the impact of AI technology on carbon emissions, with a significant "inverted U-shaped" relationship between AI technology and carbon emissions in the eastern and western regions. In the central region, AI technology has not achieved carbon emission reduction, but had a significant contribution to carbon emissions.#br#The marginal contributions of this paper are as follows. First, from the perspective of research, this paper examines the effect of AI technology on carbon emissions in a more systematic way from the perspective of AI technology for the first time. Second, in terms of the research methodology, AI patent data are collected manually to measure the level of AI technology in each region, and a panel semi-parametric model and a nonlinear mediated effects model are used. Third, from the research findings, it is found that there is an "inverted U-shaped" relationship between AI technology and carbon emissions, and there is regional heterogeneity in the effect of AI. This provides empirical support for the development of differentiated AI carbon emission reduction strategies.#br#