人工智能与包容性增长——来自全球工业机器人使用的证据

Artificial Intelligence and Inclusive Growth:Evidence from International Robot Data

人工智能是否引发收入差距扩大,关系到社会不平等的抑制与包容性增长的实现。本文采用产业层面的世界工业机器人安装数据,通过划分高中低三个技能层次的劳动者,考察了人工智能应用对各国产业内收入差距的影响、作用渠道及其产业间的溢出效应。研究发现:在岗位更迭效应和生产率效应的作用下,人工智能总体上促进了产业内的包容性增长,缩小了不同阶层劳动者的劳动收入差距;尤其是经济发展水平较高的国家、产业智能化发展期和处于上升期的产业更容易享受到岗位更迭效应与生产率效应带来的红利。但是,在各产业生产日趋网络化的条件下,人工智能应用也带来了产业间的负面溢出效应,加剧了其他产业的收入差距。本文结论有助于更加精准地帮助中低收入群体提高其收入水平,有计划地逐步推进产业智能化。

China has been striving to achieve the goal of common prosperity.As the industrial intelligence grows as an increasingly important driving force and source of economic growth in the modern society,whether and to what extent the poor people can benefit from the wide application of artificial intelligence greatly affect the balance between economic growth and income equality.The notion of inclusive growth,aiming at promoting economic growth,narrowing income gaps,and improving the income of the poor people,attracts increasing attention from both academia and policy-makers.The existing literature has not reached a consensus on the impact of technological progress on inclusive growth.In the early 20 th century,John Maynard Keynes pointed out that human beings would face“technological unemployment”.However,some scholars argue that artificial intelligence has the potential to increase productivity,create new jobs,at the same time,reduce income gaps and achieve inclusive growth.To address these issues,we build a theoretical framework and provide updated and comprehensive evidence for the impact of artificial intelligence on the inclusive growth based on the Industrial Robots report by International Federation of Robotics(IFR)and the industry-level data by the World Input-output Database.Our empirical estimations show that artificial intelligence has overall promoted the inclusive growth within industries;it increases the share of labor income and reduces the income gap between the low-skilled and highly-skilled workers and between the medium-skilled and highly-skilled workers.Behind this finding are the high productivity and displacement effect dominated by job creation,which promotes inclusive growth.We emphasize that this positive impact shows significant heterogeneity.Artificial intelligence tends to promote inclusive growth in countries with advanced development level,as well as in industries with better established intelligent development and growth prospects,bring about displacement effect and productivity effect.Similar impacts do not occur in the case of a low productivity level,and where job substitution dominates the displacement effect follows the development of artificial intelligence.In addition to the above intra-industry effects,our analysis further shows that artificial intelligence generates a negative inter-industry spillover effect.Endogenously mediated by labor mobility,productivity,and innovation rent,the increase of intelligence in other industries leads to an increase in the income gap in own industry.This paper contributes to the existing studies in three dimensions.First,our analysis is not restricted to the high-income or middle-income countries as in most prior studies.We make use of the Industrial Robots Report by IFR and broaden the sample to 55 industries in 41 countries,with an attempt to obtain more general conclusions.Second,instead of assuming the exogenous applications of artificial intelligence as in most of the existing studies,for the first time,we resort to the endogenous mediation effect model to investigate the endogenous working mechanism of artificial intelligence influencing inclusive growth.Third,our analysis focuses on the balance between economic growth and income distribution.We confirm the overall positive impact of artificial intelligence on economic growth and explore the conditions under which it also promotes income distribution.At the same time,the heterogeneity study is no longer limited to the static industrial perspective.Instead,we consider the differences between industries and the dynamic cycle within industries,to discuss whether the impact of artificial intelligence on inclusive growth has cyclical characteristics and spillover effects.We also explore the inter-industry spillover effects by referring to the method of social network analysis.We expand the investigation of spillover effects from the upstream and downstream industrial chain to the production network among various industries,and discuss the impact of the intelligence level of other industries on the income distribution gap of this industry.Our findings have strong policy implications.As an important driving force of economic growth,industrial intelligence has the potential to raise the income of the poor in a large proportion,hence promoting income distribution and realizing inclusive growth.In the era of digital transformation,blindly worrying about artificial intelligence or hastily blocking its development should not be encouraged.Being aware of the possible adverse effects of artificial intelligence on low-skilled workers,the government could avoid a one-size-fits-all pattern in the process of promoting industrial intelligence.Based on the initial distribution of income within and across industries/areas,policy-makers can be more flexible in creating employment opportunities for low-skilled workers and supporting industries of low productivity,so as to promote joint growth without income gap being widened.