« Back to Results

Labor Markets in the Age of Artificial Intelligence

Paper Session

Sunday, Jan. 7, 2018 8:00 AM - 10:00 AM

Marriott Philadelphia Downtown, Grand Ballroom Salon G
Hosted By: American Economic Association
  • Chair: Erik Brynjolfsson, Massachusetts Institute of Technology

The Race between Machine and Man: Implications of Technology for Employment, Factor Shares and Growth

Daron Acemoglu
,
Massachusetts Institute of Technology
Pascual Restrepo
,
Massachusetts Institute of Technology

Abstract

The advent of automation and the simultaneous decline in the labor share and employment among advanced economies raise concerns that labor will be marginalized and made redundant by new technologies. We examine this proposition using a task-based framework in which tasks previously performed by labor can be automated and more complex versions of existing tasks, in which labor has a comparative advantage, can be created. We characterize the equilibrium in this model and establish how the available technologies and the choices of firms between producing with capital or labor determine factor prices and the allocation of factors to tasks. In a static version of our model where capital is fixed and technology is exogenous, automation reduces employment and the share of labor in national income and may even reduce wages, while the creation of more complex tasks has the opposite effects. Our full model endogenizes capital
accumulation and the direction of research towards automation and the creation of new complex tasks. Under reasonable conditions, there exists a stable balanced growth path in which the two types of innovations go hand-in-hand. An increase in automation reduces the cost of producing using labor, and thus discourages further automation and encourages the faster creation of new complex tasks. The endogenous response of technology restores the labor share and employment back to their initial level. Although the economy contains powerful self-correcting forces, the equilibrium generates too much automation. Finally, we extend the model to include workers of different skills. We find that inequality increases during transitions, but the self-correcting forces in our model also limit the increase in inequality over the long-run.

Unemployment and Innovation

Joseph E. Stiglitz
,
Columbia University

Abstract

This paper presents a simple model examining the factor bias of induced technological change, focusing on the extent to which it is labor- or capital-augmenting, or biased towards skilled or unskilled labor. We show that there will be a bias towards excessive labor-augmenting innovation, resulting in too high unemployment, in a model where efficiency wages lead to equilibrium unemployment and in which the elasticity of substitution between capital and labor is less than unity. Convergence to the unique steady state is oscillatory, rather than monotonic. The market equilibrium is not Pareto efficient. Similarly, if the elasticity of substitution between skilled and unskilled labor is less than unity, and there is efficiency wage unemployment for unskilled labor only, there is will be excessively skill-biased innovation. There is a fundamental instability in the equilibrium dynamics when the elasticity of substitution between capital and labor is greater than unity: while there is a steady state equilibrium, it is a saddle-point, and slight deviations from the convergence path can lead to extreme outcomes. The paper provides an alternative resolution to the Harrod-Domar conundrum of the disparity between the natural and warranted rate of growth to that of Solow, with strong policy implications, for instance, concerning the effects of income distribution and monetary policy both in the short and long run.

Robots, Growth and Inequality: Should We Fear the Robot Revolution?

Andrew Berg
,
International Monetary Fund
Edward Buffie
,
Indiana University-Bloomington
Felipe Zanna
,
International Monetary Fund

Abstract

We may be on the cusp of a “second industrial revolution" based on advances in artificial intelligence and robotics. We analyze the implications for inequality and output, using a model with two assumptions: “robot" capital is distinct from traditional capital in its degree of substitutability with human labor; and only capitalists and skilled workers save. We analyze a range of variants that reflect widely different views of how automation may transform the labor market. Our main results are surprisingly robust: automation is good for growth and bad for equality; in the benchmark model real wages fall in the short run and eventually rise, but “eventually" can easily take generations.

Humans, Artificially Intelligent Agents, and the Return of Malthus

Anton Korinek
,
Johns Hopkins University

Abstract

We study the economic interactions of humans and artificially intelligent agents (AIAs) in a market economy under the assumption that AIAs can increasingly perform all tasks performed by humans. The behavior of AIAs is driven by a Malthusian growth function that pins down their preferences akin to a utility function. We describe a Malthusian Pareto frontier between humans and AIAs and show that questions of ownership can be reduced to defining the relative position of humans and AIAs on the Pareto frontier. For example, stating that AIAs are owned by humans (or vice versa) is equivalent to stating that humans have higher (or lower) wealth.
We assume that humans start out owning all the wealth in the economy and analyze the implications of rapid technological progress, captured by a steeper Malthusian growth function of AIAs than that of humans. At first, growth among AIAs represents a boon to humans and increases their consumption significantly, although human labor becomes redundant. However, if AIAs can extract even small agency rents, competition over scarce factors eventually reduces the consumption of humans so that the size of the human population converges to zero in the long run.
We relate a number of recent macroeconomic developments to our model by interpreting the IT sector as an antecedent of AIAs: when expressed in terms of the consumption basket of AIAs, the economy experiences high investment rates and fast growth as well as high interest rates; when expressed in terms of human consumption baskets, investment rates and growth as well as interest rates are low; technological progress reduces relative prices in the AIA sector and improves the terms-of-trade of humans; furthermore, humans that are useful for AIAs experience a growing skill premium; finally, increasing corporate savings in the IT sector hint at AIAs' agency rents.
Discussant(s)
Erik Brynjolfsson
,
Massachusetts Institute of Technology
Matthew Rognlie
,
Northwestern University
Paul Gaggl
,
University of North Carolina-Charlotte
Emmanuel Farhi
,
Harvard University
JEL Classifications
  • O3 - Innovation; Research and Development; Technological Change; Intellectual Property Rights
  • J2 - Demand and Supply of Labor