The Economics of Artificial Intelligence
Page 65
Challenges for Development, edited by Mario Cimoli, Giovanni Dosi, Keith E.
Maskus, Ruth L. Okediji, Jerome H. Reichman, and Joseph E. Stiglitz, 1– 53.
Oxford: Oxford University Press.
Drandakis, Emmanuel, and Edmund Phelps. 1965. “A Model of Induced Invention,
Growth and Distribution.” Economic Journal 76:823– 40.
Fehr, Ernst, and Klaus M. Schmidt. 2003. “Theories of Fairness and Reciprocity—
Evidence and Economic Applications.” In Advances in Economics and Economet-
rics, Econometric Society Monographs, Eighth World Congress, vol. 1, edited
by Mathias Dewatripont, Lars Peter Hansen, and Stephen J Turnovsky, 208– 57.
Cambridge: Cambridge University Press.
Financial Crisis Inquiry Commission. 2011. “The Financial Crisis Inquiry Report.”
Final Report of the National Commission. January. http:// www .gpoaccess .gov
/ fcic/ fcic .pdf.
Frey, Carl Benedikt, and Michael A. Osborne. 2017. “The Future of Employment:
How Susceptible Are Jobs to Computerisation?” Technological Forecasting and
Social Change 114:254– 80.
Geanakoplos, John, and Herakles Polemarchakis. 1986. “Existence, Regularity, and
Constrained Suboptimality of Competitive Allocations When the Asset Market
Is Incomplete.” In Uncertainty, Information and Communication: Essays in Honor
of KJ Arrow, edited by W. Heller, R. Starr, and D. Starrett, 65– 96. Cambridge: Cambridge University Press.
Gordon, Robert. 2016. The Rise and Fall of American Growth: The U.S. Standard of
Living since the Civil War. Princeton, NJ: Princeton University Press.
Greenwald, Bruce, and Joseph E. Stiglitz. 1986. “Externalities in Economics with
Imperfect Information and Incomplete Markets.” Quarterly Journal of Economics
101 (2): 229– 64.
Groshen, Erica L., Brian C. Moyer, Ana M. Aizcorbe, Ralph Bradley, and David M.
Friedman. 2017. “How Government Statistics Adjust for Potential Biases from
AI and Its Implications for Income Distribution and Unemployment 389
Quality Change and New Goods in an Age of Digital Technologies: A View from
the Trenches.” Journal of Economic Perspectives 31 (2): 187– 210.
Guzman, Martin, and Joseph E. Stiglitz. 2016a. “Pseudo- Wealth and Consumption
Fluctuations.” NBER Working Paper no. 22838, Cambridge, MA.
Guzman, Martin, and Joseph E. Stiglitz. 2016b. “A Theory of Pseudo- Wealth.” In
Contemporary Issues in Macroeconomics: Lessons from The Crisis and Beyond,
edited by Joseph E. Stiglitz and Martin Guzman. IEA Conference Volume no.
155-II. Basingstoke, UK: Palgrave Macmillan.
Hanson, Robin. 2016. The Age of Em. Oxford: Oxford University Press.
Harari, Yuval N. 2017. Homo Deus: A Brief History of Tomorrow. New York: Harper.
Hicks, John. 1932. The Theory of Wages. London: Macmillan.
Kennedy, Charles. 1964. “Induced Bias in Innovation and the Theory of Distribu-
tion.” Economic Journal LXXIV:541– 47.
Keynes, John Maynard. 1932. “Economic Possibilities for our Grandchildren.” In
Essays in Persuasion, 358– 73. San Diego, CA: Harcourt Brace.
Korinek, Anton. 2016. “Currency Wars or Effi
cient Spillovers? A General Theory
of International Policy Cooperation.” NBER Working Paper no. 23004, Cam-
bridge, MA.
———. 2017. “The Rise of Artifi cially Intelligent Agents.” Working paper, Johns
Hopkins University.
Korinek, Anton, and Ding Xuan Ng. 2018. “Digitization and the Macroeconomics of
Superstars.” Working paper, Johns Hopkins University and University of Virginia.
Korinek, Anton, and Joseph E. Stiglitz. 2017. “Worker- Replacing Technological
Progress.” NBER Working Paper no. 24174, Cambridge, MA.
Kurzweil, Ray. 2005. The Singularity Is Near: When Humans Transcend Biology.
New York: Viking.
Lipsey, Richard, and Kelvin Lancaster. 1956. “The General Theory of Second Best.”
Review of Economic Studies 24 (1): 11– 32.
Malthus, Thomas Robert. 1798. An Essay on the Principle of Population. Project
Gutenberg.
Manning, Alan. 2011. “Imperfect Competition in Labour Markets.” In Handbook of
Labor Economics, vol. 4, edited by O. Ashenfelter and D. Card. North- Holland:
Amsterdam.
Meade, James E. 1955. Trade and Welfare. Oxford: Oxford University Press.
Moore, Gordon E. 1965. “Cramming More Components onto Integrated Circuits.”
Electronics 38 (8): 114:ff .
Mortensen, Dale T., and Christopher A. Pissarides. 1994. “Job Creation and Job
Destruction in the Theory of Unemployment.” Review of Economic Studies 61
(3): 397– 415.
———. 1998. “Technological Progress, Job Creation, and Job Destruction.” Review
of Economic Dynamics 1 (4): 733– 53.
Newbery, David, and Joseph E. Stiglitz. 1984. “Pareto Inferior Trade.” Review of
Economic Studies 51 (1): 1– 12.
Omohundro, Stephen M. 2008. “The Basic AI drives.” In Artifi cial General Intel-
ligence 2008: Proceedings of the First AGI Conference, edited by Pei Wang, Ben
Goertzel, and Stan Franklin, 483– 92. Amsterdam: IOS.
Piketty, Thomas, Emmanuel Saez, and Stefanie Stantcheva. 2014. “Optimal Taxa-
tion of Top Labor Incomes: A Tale of Three Elasticities.” American Economic
Journal: Economic Policy 6 (1): 230– 71.
Restrepo, Pascual. 2015. “Skill Mismatch and Structural Unemployment.” Working
paper, Massachusetts Institute of Technology.
Samuelson, Paul. 1965. “A Theory of Induced Innovations along Kennedy- Weisacker
Lines.” Review of Economics and Statistics XLVII:444– 64.
390 Anton Korinek and Joseph E. Stiglitz
Schmitt, John. 2013. Why Does the Minimum Wage Have No Discernible Eff ect on
Employment? Washington, DC: Center for Economic and Policy Research.
Shapiro, Carl, and Joseph E. Stiglitz. 1984. “Equilibrium Unemployment as a
Worker Discipline Device.” American Economic Review 74 (3): 433– 44.
Simsek, Alp. 2013. “Speculation and Risk Sharing with New Financial Assets.”
Quarterly Journal of Economics 128 (3): 1365– 96.
Solow, Robert. 1987. “We’d Better Watch Out.” New York Times Book Review,
July 12, 36.
Stiglitz, Joseph E. 1969. “Distribution of Income and Wealth among Individuals.”
Econometrica 37 (3): 382– 97.
———. 1976. “The Effi
ciency Wage Hypothesis, Surplus Labour and the Distribu-
tion of Income in LDCs.” Oxford Economic Papers 28:185– 207.
———. 1987a. “On the Microeconomics of Technical Progress.” In Technology
Generation in Latin American Manufacturing Industries, edited by Jorge M. Katz, 56– 77. New York: St. Martin’s Press.
———. 1987b. “Technological Change, Sunk Costs, and Competition.” Brookings
Papers on Economic Activity 3:883– 947.
———.2006. “Samuelson and the Factor Bias of Technological Change.” Samuel-
sonian Economics and the Twenty- First Century, edited by M. Szenberg et al., 235–
51. New York: Oxford University Press.
———. 2008. “The Economic Foundations of Intellectual Property.” Duke Law
Journal 57 (6): 1693– 724.
———. 2014a. “Intellectual Property Rights, the Pool of Knowledge, and Innova-
tion.” NBER
Working Paper no. 20014, Cambridge, MA.
———. 2014b. “Unemployment and Innovation.” NBER Working Paper no. 20670,
Cambridge, MA.
———. 2014c. “Tapping the Brakes: Are Less Active Markets Safer and Better for
the Economy?” Presentation at the Federal Reserve Bank of Atlanta 2014 Finan-
cial Markets Conference.
———. 2017. “Pareto Effi
cient Taxation and Expenditures: Pre- and Re-
distribution.” NBER Working Paper no. 23892, Cambridge, MA.
Stiglitz, Joseph E., and Bruce Greenwald 2015. Creating a Learning Society: A New
Approach to Growth, Development, and Social Progress. New York: Columbia
University Press.
Stiglitz, Joseph E., with Nell Abernathy, Adam Hersh, Susan Holmberg, and Mike
Konczal. 2015. Rewriting the Rules of the American Economy, A Roosevelt Insti-
tute Book. New York: W. W. Norton.
Stiglitz, Joseph E., and Andrew Weiss. 1981. “Credit Rationing in Markets with
Imperfect Information.” American Economic Review 71 (3): 393– 410.
Tegmark, Max. 2017. Life 3.0: Being Human in the Age of Artifi cial Intelligence.
New York: Knopf.
Turing, Alan M. 1950. “Computing Machinery and Intelligence.” Mind 59 (236): 433– 60.
United Nations Department of Economic and Social Aff airs (UN). 2015. “United
Nations World Population Prospects: 2015 Revision.” https:// esa.un .org/ unpd
/ wpp/ publications/ fi les/ key_fi ndings_wpp_2015 .pdf.
Vinge, Vernor. 1993. “The Coming Technological Singularity: How to Survive in the
Post- Human Era.” In Proc. Vision 21: Interdisciplinary Science and Engineering in
the Era of Cyberspace, 11– 22. NASA: Lewis Research Center.
von Weizacker, C. C. 1966. “Tentative Notes on a Two- Sector Model with Induced
Technical Progress.” Review of Economic Studies 33:245– 51.
Williams, Heidi. 2010. “Intellectual Property Rights and Innovation: Evidence from
the Human Genome.” NBER Working Paper no. 16213, Cambridge, MA.
15
Neglected Open Questions in the
Economics of Artifi cial Intelligence
Tyler Cowen
Many recent writings consider artifi cial intelligence (AI), or more broadly
“smart software,” as a transformative technology. Commonly, these writings
focus on the substitution of capital for labor and the attendant domestic
labor market eff ects. Without meaning to downplay the importance of that
topic, I’d like to focus our attention on some other aspects of how artifi cial
intelligence might aff ect our society.
15.1 The Distribution of Consumer Surplus
Most analyses of automation focus on the production function, but the
new and cheaper outputs resulting from automation have distributional
eff ects as well. For instance, the Industrial Revolution made food cheaper
and more reliable in supply, in addition to mechanizing jobs in the factory
and in the fi elds. A new, larger, cheaper and more diverse book market was
created, and so on. Artifi cial intelligence, in turn, holds out the prospect of
lowering prices for the outputs that can be produced by the next generation
of automation. Imagine education and manufactured goods being much
cheaper because we produced them using a greater dose of smart software.
The upshot is that even if a robot puts you out of a job or lowers your pay,
there will be some recompense on the consumer side. Internet goods such
as Facebook already constitute a signifi cant part of individuals’ time alloca-
tion, and of course they are free or very cheap at the relevant margin.
It’s worth thinking about whether the new AI- enabled outputs will be pro-
Tyler Cowen is professor of economics at George Mason University.
For acknowledgments, sources of research support, and disclosure of the author’s material fi nancial relationships, if any, please see http:// www .nber .org/ chapters/ c14032.ack.
391
392 Tyler Cowen
duced at constant, increasing, or declining cost. Usually software- intensive
goods tend to be produced at declining cost; namely, there is an upfront
investment in the software, but at the margin additional copies are quite
cheap or possibly free.
The declining cost scenario seems to have some optimistic properties. If
the marginal cost is zero or near- zero, in the longer run the output price
should fall considerably. In some cases, such as with social networks, the
price may be zero to begin with, or perhaps negative to encourage people
to join the network. Once we consider these consumption side eff ects, the
distributional implication of an AI revolution could be more egalitarian
than the job displacement eff ects alone would indicate.
For instance, consider the role of smartphones and cell phones in Africa
today. These items have a relatively low marginal cost, and they are sold
in Africa quite cheaply. They have transformed some sectors of African
economies by making it much easier to manage businesses, and they also
allow Africans the pleasure of communicating with each other more easily.
The substitution of labor for capital in smartphone manufacturing hasn’t
impacted African economies much at all because Africa is not a major part
of the supply chain. The more that tech production is clustered, the more
that the consumption eff ects will be the major eff ects for many parts of the
world.
These distribution eff ects may be less egalitarian if hardware rather than
software is the constraint for the next generation of AI. Hardware is more
likely to exhibit constant or rising costs, and that makes it more diffi
cult
for suppliers to charge lower prices to poorer buyers. You might think it is
obvious that future productivity gains will come in the software area—and
maybe so—but the very best smart phones, such as iPhones, also embody
signifi cant innovations in the areas of materials. A truly potent AI device
might require portable hardware at signifi cant cost. At this point we don’t
know, but it would be unwise to assume that future innovations will be
software- intensive to the same extent that recent innovations have been.
If future AI innovations lead to very low consumer prices, this may aff ect
our policy recommendations. Often analysts who are worried about automa-
tion call for better education and job training. Those may still be good ideas,
but another approach can pay off as well. To the extent productivity is very
high and prices are very low, it may suffi
ce for workers to own some capital or
natural resources. That is, wealth can serve as a substitute for income, given
the extremely high purchasing power resulting from the low prices. Giving
everyone some land, a birthright grant or shares in a sovereign wealth fund
are options to consider, on top of whatever changes might be made to edu-
cation and labor markets.
Perhaps counterintuitively, the economics of natural resources would
become signifi cantly more relevant in such a world. The scarcity of labor
would matter much less, and of course robots could be used to make more
Neglected Open Questions in the Economics of Artifi cial Intelligence 393
robots. Y
ou might even imagine software programs generating new prod-
ucts and ideas, and organizing their implementation. What would, in fact,
constrain production? The answer is energy and possibly land. As scarce
inputs, land and energy would determine which economies would do well
and which not so well. In such a world the returns to education could be
very low rather than very high.
An alternative possibility for the new scarce resource might be institu-
tions to encourage AI- led production, such as maximally secure property
rights. In that case, public choice factors would become a more signifi cant
determinant of national and regional outcomes. If “good government” is a
public good of sorts, that would benefi t nations and regions with especially
eff ective norms for governance, for instance Singapore.
15.2 International
Eff ects of an AI Revolution
Information technology also interacts with international trade. One eff ect
of smart software is to enable more factor price equalization. It helps suc-
cessful businesses become larger and also branch out internationally; for
instance, it would be harder for Apple to fi nish off the iPhone in China if
it only had the communications technologies of a few decades ago. These
days, company leadership can manage an international empire by cell phone,
email, and other technologies, and arguably that has led to higher investment
in Chinese workers and lower investment in American and other developed
country workers, especially at the lower- skilled end of the distribution.
That said, if you imagine artifi cial intelligence and other technologies
progressing further, wage diff erentials might cease to be a reason to locate
abroad at all. Why should the wage diff erential matter if the company is
hardly employing any labor? As a result, there might be a reshoring of
American or Western European manufacturing.
This could boost the demand for janitors here in the United States and
also increase their wages, even though the number of such janitors might be
small. Possibly the big income distribution eff ect is that artifi cial intelligence
will be much worse for the poorer countries that can no longer industrial-
ize through wage diff erentials; Dani Rodrik has labeled this phenomenon
“premature deindustrialization.” At the same time, AI may be just fi ne for