The Korean Economic Forum
AI, Employment, Economic Growth, and Inequality: Recent Literature Survey and Policy Implications
Young Sik Kim (Seoul National University)Year 2019Vol. 12No. 3
Abstract
The main findings and policy implications of recent studies concerning theeffects of artificial intelligence (AI) on employment, economic growth andinequality are as follows. First, the expansion of automated tasks due to theintroduction of AI technology has the effect of reducing labor demand, but thecreation of new tasks can lead to an increase in labor demand, wage level andlabor income share through re-instatement effect that strengthens theproductivity effect of automation. Second, as more sectors are automated, boththe share of goods produced in the automated sectors and capital income shareincrease. However, the price of automated commodities whose productionincrease rapidly decreases, which results in a drop in their GDP share due tothe “cost disease” (Baumol, 1967). In the long run, if a constant fraction of thenon-automated sectors becomes automated annually and the economyasymptotically converges to full automation, a balanced growth path is attainedin which both the labor-augmenting technology level and GDP per capita growat a constant rate and factor income shares of capital and labor are constant,respectively. Finally, in the real world where the market is incomplete andredistribution involves nontrivial cost, AI-related technological innovations candeepen inequality by causing innovators to possess all the surpluses. Therefore,in order to increase social safety nets and reduce inequality, active redistributionmeasures have been proposed such as various tax instruments, institutionalchanges that bring more favorable market allocation to workers, and universalbasic income.}