Book Read Free

The Economics of Artificial Intelligence

Page 31

by Ajay Agrawal


  sance Man’: Is Innovation Getting Harder.” Review of Economics and Statistics

  76 (1): 283– 317.

  174 Ajay Agrawal, John McHale, and Alexander Oettl

  Jones, Charles. 1995. “R&D-Based Models of Economic Growth.” Journal of Po-

  litical Economy 103 (4): 759– 84.

  ———. 2005. “Growth and Ideas.” Handbook of Economic Growth, vol. 1B, edited

  by Phillipe Aghion and Steven Durlauf. Amsterdam: Elsevier.

  Kauff man, Stuart. 1993. The Origins of Order. Oxford: Oxford University Press.

  LeCun, Yann, Yoshua Bengio, and Geoff rey Hinton. 2015. “Deep Learning.” Nature

  521:436– 44.

  Levinthal, Daniel. 1997. “Adaptation on Rugged Landscapes.” Management Science

  43:934– 50.

  Mamoshina, Polina, Armando Vieira, Evgeny Putin, and Alex Zhavoronkov. 2016.

  “Applications of Deep Learning in Biomedicine.” Molecular Pharmaceutics

  13:1445– 54.

  Mokyr, Joel. 2002. The Gifts of Athena: Historical Origins of the Knowledge Economy.

  Princeton, NJ: Princeton University Press.

  ———. 2017. “The Past and Future of Innovation: Some Lessons from Economic

  History.” Paper presented at the NBER Conference on Research Issues in Artifi -

  cial Intelligence, Toronto, Sept. 2017.

  Mullainathan, Sendhil, and Jann Spiess. 2017. “Machine Learning: An Applied

  Econometric Approach.” Journal of Economic Perspectives 31 (2): 87– 106.

  National Science and Technology Council. 2011. “Materials Genome Initiative for

  Global Competitiveness.” Washington, DC.

  Nelson, Richard, and Sidney Winter. 1982. An Evolutionary Theory of Economic

  Change. Cambridge, MA: Harvard University Press.

  Nielsen, Michael. 2012. Reinventing Discovery: The New Era of Networked Science.

  Princeton, NJ: Princeton University Press.

  Nordhaus, William. 2015. “Are We Approaching an Economic Singularity? Informa-

  tion Technology and the Future of Economic Growth.” NBER Working Paper

  no. 21547, Cambridge, MA.

  Rivkin, Jan. 2000. “Imitation of Complex Strategies.” Management Science 46:824– 44.

  Romer, Paul. 1990. “Endogenous Technical Change.” Journal of Political Economy

  94: S71– 102.

  ———. 1993. “Two Strategies for Economic Development: Using and Producing

  Ideas.” In Proceedings of the World Bank Annual Conference on Development Eco-

  nomics, 1992. Washington, DC: The World Bank.

  ———. 2008. “Economic Growth.” In The Concise Encyclopaedia of Economics,

  Library of Economic Liberty. http:// www .econlib .org/ library/ Enc/ Economic

  Growth .html.

  Rosenberg, Nathan. 1998. “Chemical Engineering as a General Purpose Tech-

  nology.” In General Purpose Technologies and Economic Growth, edited by Elhanan Helpman. Cambridge, MA: MIT Press.

  Weitzman, Martin. 1998. “Recombinant Growth.” Quarterly Journal of Economics

  113:331– 60.

  Zeira, Joseph. 1998. “Workers, Machines, and Economic Growth.” Quarterly Journal

  of Economics 113 (4): 1091– 117.

  6

  Artifi cial Intelligence

  as the Next GPT

  A Political- Economy Perspective

  Manuel Trajtenberg

  6.1 Introduction

  Artifi cial intelligence (AI) and related technologies are being heralded

  as “the next big thing,” one that promises to revolutionize many areas of

  economic activity and thus to have a profound impact on economic growth.

  However, the rise of AI coincides with a recent wave of pessimism in terms

  of productivity growth, expressed forcefully by prominent economists such

  as Larry Summers (2016), and more thoroughly by Robert Gordon (2016).

  Side by side with the gloom, the new “technology enthusiasts” envision

  a not- too- distant future in which AI will displace most ( all?) human occu-

  pations while unleashing tremendous gains in productivity. This view poses

  once again disturbing questions about the future of employment, the distri-

  butional consequences of mass displacement, and so forth.

  Nobody holds the crystal ball, hence rather than arguing about the inscru-

  table future, it is at least as important to inquire into what we can learn from

  history regarding episodes like this, that is, the appearance of a major new

  technology that is posed to have profound economic implications. Of course,

  the future is never a replay of the past, but it may provide a useful benchmark

  against which to assess the unfolding of the new technology.

  Mokyr (2017) sounds a cautionary note in that regard: ever since the

  dawn of the Industrial Revolution in the late eighteenth century, both the

  pessimists and the enthusiasts have almost invariably been proven wrong.

  Manuel Trajtenberg is a professor in the Eitan Berglas School of Economics at Tel Aviv University and a research associate of the National Bureau of Economic Research.

  This is a follow-up to my discussion at this conference of Joel Mokyr’s paper “The Past and the Future of Innovation: some lessons from Economic History.” 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/ c14025.ack.

  175

  176 Manuel Trajtenberg

  Moreover, Mokyr dismantles with solid historical and present day evidence

  Gordon’s claim that technological advance is bound to slow down in a deter-

  ministic fashion (in particular, the claim that “all low- hanging scientifi c and

  technological fruit has already been picked”).

  However, nothing can be taken for granted—as Mokyr skillfully describes,

  institutions (including government policies) may play a key role enabling or

  retarding innovation. This is precisely the focus here: given that AI is poised

  to emerge as a powerful technological force, I discuss ways to mitigate the

  almost unavoidable ensuing disruption, and enhance AI’s vast benign poten-

  tial. This is particularly important in present times, in view of political-

  economic considerations that were mostly absent in previous historical epi-

  sodes associated with the arrival of new general purpose technologies.

  6.2 Is This Time Diff erent? The Political Economy

  of Technological Disruptions

  The presumption here, well argued in other papers in this conference,1 is

  that AI has the potential of becoming a general purpose technology (GPT)

  in the foreseeable future, 2 thus bringing about a wave of complementary

  innovations in a wide and ever- expanding range of applications sectors.

  Such sweeping transformative processes always result in widespread eco-

  nomic disruption, with concomitant winners and losers.

  The “winners” are primarily those associated with the emerging GPT

  sector itself, and those that are at the forefront of the deployment of the

  GTP in the main applications sectors. They tend to be young, entrepre-

  neurial, and equipped both with the technical knowledge and the skills that

  are made relevant by the new GPT. The labor force composition of Silicon

  Valley off ers a grand view of who are the winners in the present informa-

  tion and communication technologies (ICT)/ internet era. There are further

  winners in those sectors that are ancillary to the core GTP circle, be it in

  services
that directly benefi t from the growth of the GPT (e.g., the venture

  capital (VC) industry, patent lawyers, designers, etc.), or in others that just

  ride on the localized boom (e.g., upscale restaurants and entertainment,

  gyms, tourism, etc.).

  The “losers” are mostly those employed in sectors that structurally cannot

  benefi t from the unfolding GPT (“laggards”), and those in industries where

  the adoption of the new GPT renders many existing competencies and skills

  obsolete, thus bringing about massive layoff s. They tend to be middle- aged,

  have lower than average educational levels, and reside in areas that do not

  have much diversifi ed sources of employment.

  As economists, we tend to view the big sweep of economic growth since

  1. See Cockburn, Henderson, and Stern (chapter 5, this volume).

  2. See Bresnahan and Trajtenberg (1995).

  AI as the Next GPT: A Political- Economy Perspective 177

  the Industrial Revolution as the very embodiment of the “Idea of Progress”

  (as conceived in the Enlightenment), and hence the rate of growth of gross

  domestic product (GDP) as an unequivocal uptick in the welfare of society

  as a whole. Sure, we do acknowledge that there are distributional conse-

  quences, and sure, ever since Pareto we know that we are not allowed to

  “sum-up utilities” (and thus the “minuses” of losers do not cancel out with

  the “pluses” of winners). But those half- hearted qualifi cations become just

  lip service—the truth is that we rarely dwell into the balance of winners and

  losers, and in particular we do not pay much attention to the later. Para-

  phrasing the well- known dictum of Isaac Newton, we may say that

  We enjoy today higher standards of living because we are standing on the

  broken backs of those that paved the way for technological progress, but

  did not live long enough to benefi t from it.

  Partly in response to these inequities, the post- World War II era saw the

  creation of the welfare state, including unemployment insurance, transfers

  to the disadvantaged, some form of health insurance, retraining programs,

  and so forth. These “safety nets” were supposed to provide a reasonable

  palliative to “losers,” but the truth is that we still do not have eff ective mecha-

  nisms to prevent or ameliorate the costs of major technologically induced

  transformations.3 Moreover, existing safety nets will quite likely fail to cope

  with the juxtaposition of two new and powerful phenomena: (a) much larger

  fl ows of GPT- displaced workers and (b) a new “great demographic transi-

  tion.” Let us examine each in turn.

  Regarding the extent of displacement: technological change always causes

  disruption, as brilliantly articulated by Schumpeter’s notion of “creative

  destruction.” Furthermore, there are infl ection points as a new GPT starts

  working its way through the economy, when in relative short notice very

  many sectors, competencies, and skills became laggards and obsolete.

  However, as clearly envisioned in this conference, AI in its various incar-

  nations seems to go much further, in that it has the potential to replace a

  very wide swath of human occupations. Many argued forcefully that there

  are no occupations that cannot be eventually replaced by AI, and that the vast majority of present occupations will indeed vanish within a generation.

  The consensual view seems to be that a large proportion of employment

  as we know it today will give way to smart machines, and therefore that x

  percent of workers will be displaced, whereby x is thought to be signifi cantly larger than in previous GPTs. At the same time, the extent to which new, presently unforeseeable occupations may arise (denote them y percent) seems

  to be constrained by the very nature of AI: presumably AI will be able to

  3. Typically, these safety nets function reasonably well when dealing with the consequences of not- too- pronounced business cycles or with small, temporarily deprived groups of the population. Not so when there are major structural transformations or when the underlying conditions that led to welfare dependency become permanent.

  178 Manuel Trajtenberg

  perform most of the new tasks, and hence they will not constitute a good

  enough counterbalance to the disappearing jobs, as has been the case in the

  past. The prevailing view is then that the net displacement of employment

  ( x – y) will turn out to be signifi cantly larger for AI than in previous epi-

  sodes of technological disruption, posing a serious challenge to traditional

  economic policies.

  The second part of the challenge entails a steep drop in birth rates together

  with the extension of life expectancy (which has been steadily growing for

  well over a century). These powerful demographic forces have resulted in

  aging populations, with the concomitant increase in the dependency ratio

  and the looming threat on the long- term viability of the pension system.

  Notice that life expectancy is now increasing well past the retirement age, so

  that a typical person in her fi fties contemplates a further stretch of twenty-

  fi ve to thirty years of life. Thus, the prospect of being permanently laid off

  at that stage in life has dire consequences for the displaced individual as

  much as for society as a whole.

  The joint eff ect of a large infl ux of displaced workers at the seemingly

  unique infl ection point posed by AI, together with their longer life expec-

  tancy, may thus create a formidable challenge that even the most advanced

  welfare state will be hard pressed to cope with. Put diff erently, we cannot

  aff ord to have many more, and longer- lived, unemployed or underemployed

  people. This is what is at stake with the advent of AI.

  There is yet another signifi cant development that magnifi es the challenge,

  and that is the democratization of expectations. The growth in income per

  capita involves not only a rise in material standards, but in other no less

  important dimensions of well- being, including reduced uncertainty and a

  concomitant heightened sense of control over our own lives, which entails

  also the expectation of having a voice in processes that aff ect us (Hirschman

  1970). Not by coincidence, economic growth and expanding democracy have

  more often than not gone hand in hand within, as well as across, countries.

  The Luddites of the early nineteenth century surely had their voice heard,

  as did their like- minded emulators over the following decades. However, they

  could hardly expect to make a dent on their fate: democracy was still highly

  limited and living standards still very low for the vast majority, so that most

  people were just consumed by the need to provide for their basic needs.

  Much has changed since, and nowadays virtually every individual in

  advanced western countries has come to expect to be entitled, at least in

  principle, to full participation in every realm of society: the political, the

  economic, and the cultural. The expectation is not just to vote in periodic

  elections, but to have an infl uence via “participatory democracy”; not just

  to hold a job, but to partake in the benefi ts of economic growth—this is

  what constitutes “the democratization of expectations.”

  We claim that in such co
ntext it has become much harder to have some

  (many?) bear the costs of technological disruption (the losers), while others

  AI as the Next GPT: A Political- Economy Perspective 179

  reap the benefi ts (the winners). Moreover, the losers have become much

  more skeptical of the vague promise that eventually the benefi ts will “trickle down” to them as well. With good reason: experience shows that the losers

  typically remain on the downside, even if the welfare state somehow softens

  their human costs. In advanced, democratic societies, people have become

  more impatient, more demanding of government, more intolerant of false

  promises, as well as of collective failures. Again, this should be surely con-

  sidered a highly positive by-product of the rise in living standards.

  The sharp split between winners and losers, if left to its own, may have

  serious consequences far beyond the costs for the individuals involved:

  when it coincides with the political divide, it may threaten the very fabric of

  democracy, as we have seen recently both in America and in Europe. Thus,

  if AI bursts into the scene and triggers mass displacement of workers, and

  demography plays out its fateful hand, the economy will be faced with a

  formidable dual challenge that may require a serious reassessment of policy

  options:

  • Governments may have to assume a wider responsibility for navigating

  eff ective transitions from old to new GPTs, and not just for alleviating

  some of the costs. As said above, the democratization of expectations

  will not allow just for cosmetic adjustments—the political economy of

  it will eventually force real change.

  • In so doing, governments may have to consider courses of action

  aimed inter alia at reducing signifi cantly the number of those that fall

  in between the cracks during such transitions: actual and potential los-

  ers are bound to become much less tolerant of their fate. This should be

  done not by attempting to slow down the pace of technical change (that

  would be silly and ineff ectual), but on the contrary, by making sure that

  many more can be brought to partake in it.

  6.3 From Threat to Promise: Strategies for the AI- GPT Era

  In order to meet the above- mentioned challenges, governments will have

  to design innovative strategies in the following key areas:

 

‹ Prev