Artificial intelligence and energy intensity in China's industrial sector: Effect and transmission channel

Liang Liu, Kun Yang, Hidemichi Fujii, Jun Liu

研究成果: Contribution to journalArticle査読

抄録

The continued development of artificial intelligence (AI) has changed production methods but may also pose challenges related to energy consumption; in addition, the effectiveness of AI differs across industries. Thus, to develop efficient policies, it is necessary to discuss the effect of AI adoption on energy intensity and to identify industries that are more significantly affected. Using data on industrial robots installed in 16 Chinese industrial subsectors from 2006 to 2016, this paper investigates both the effect of AI on energy intensity and the channel through which this effect is transmitted. The empirical results show, first, that boosting applications of AI can significantly reduce energy intensity by both increasing output value and reducing energy consumption, especially for energy intensities at high quantiles. Second, compared with the impacts in capital-intensive sectors (e.g., basic metals, pharmaceuticals, and cosmetics), the negative impacts of AI on energy intensity in labor-intensive sectors (e.g., textiles and paper) and technology-intensive sectors (e.g., industrial machinery and transportation equipment) are more pronounced. Finally, the impact of AI on energy intensity is primarily achieved through its facilitation of technological progress; this accounts for 78.3% of the total effect. To reduce energy intensity, the Chinese government should effectively promote the development of AI and its use in industry, especially in labor-intensive and technology-intensive sectors.

本文言語英語
ページ(範囲)276-293
ページ数18
ジャーナルEconomic Analysis and Policy
70
DOI
出版ステータス出版済み - 6 2021

All Science Journal Classification (ASJC) codes

  • Economics and Econometrics
  • Economics, Econometrics and Finance (miscellaneous)

フィンガープリント 「Artificial intelligence and energy intensity in China's industrial sector: Effect and transmission channel」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル