Professional Documents
Culture Documents
FD1223
FD1223
DECEMBER 2023
ARTIFICIAL
INTELLIGENCE
What AI means for
economics
F&D
20 42 54
The Macroeconomics of Technology for Development China’s Bumpy Path
Artificial Intelligence AI must be carefully adapted to benefit Growth slows and risks abound, but
The collective decisions we make the poor, shows research in Kenya, economic and financial collapse can
today will determine how AI affects Togo, and Sierra Leone be avoided
productivity growth, income inequality, Daniel Björkegren and Joshua Blumenstock Eswar Prasad
and industrial concentration
Erik Brynjolfsson and Gabriel Unger
46 59
AI in Practice History’s Inflation Lessons
26 Technology is reshaping the way we A study of 100 inflation shocks since
Rebalancing AI cultivate food, care for our health, and the 1970s provides valuable pointers
The drive toward automation is preserve national security for policymakers today
perilous—to support shared prosperity, Robert Horn, Jeremy Wagstaff, and Anil Ari and Lev Ratnovski
AI must complement workers, Kerry Dooley Young
not replace them
Daron Acemoglu and Simon Johnson 62
52 A Critical Matter
AI Lexicon Fragmentation of critical mineral markets
30 A brief guide to artificial would slow the shift to clean energy
Scenario Planning for an intelligence lingo Christopher Evans, Marika Santoro, and
A(G)I Future Martin Stuermer
AI may be on a trajectory to surpass
human intelligence; we should
be prepared
Anton Korinek
34
Technology’s Bifurcated Bite
Some workers will win, others will
lose as the use of artificial intelligence
grows
Andrew Berg, Chris Papageorgiou, and
Maryam Vaziri
38
AI’s Reverberations
across Finance
Financial institutions are forecast
to double their spending on AI
20
JU N C EN
by 2027
Jeff Kearns
DECEMBER 2023 1
F&D Contents
Departments
6 70 66
Kaleidoscope Straight Talk
Our view of global issues and data points Harnessing AI for Global Good
Maximizing the benefits of artificial
intelligence and managing the risks
8 will require innovative policies with
Back to Basics global reach
Artificial Intelligence’s Promise Gita Gopinath
and Peril
Generative AI is poised to unleash a
wave of creativity and productivity but 73
poses questions for humanity Book Reviews
Hervé Tourpe Preparing for the Inevitable
Risk and Resilience in the Era of Climate
Change, Vinod Thomas
10
Point of View Thinking about Inequality
Building Blocks for AI Governance Visions of Inequality from the French
Policymakers must adhere to five guiding Revolution to the End of the Cold War, 76
principles to govern AI effectively Branko Milanovic
Ian Bremmer and Mustafa Suleyman
The Flourishing Society
Fostering More Inclusive My Journeys in Economic Theory,
Democracy with AI Edmund Phelps
AI can enhance democratic institutions
by ensuring citizens’ voices are truly
heard 76
Hélène Landemore Currency Notes
On the Brink
Unlocking India’s Potential with AI Catalyzing support for the
India is on the brink of a transformation Philippine national bird could stem
that could change its economic and its rapid decline
social future Analisa R. Bala
Nandan Nilekani and Tanuj Bhojwani
TECH
Rebalancing
I N T E R N AT I O N A L M O N E TA RY F U N D
COUNTRY FOCUS
China's Bumpy
CLIMATE
A Critical
DECEMBER 2023
On the Cover
ARTIFICIAL
2 D E C E M B E R 2023
NEW
FROM THE IM
F
EDITOR-IN-CHIEF
Gita Bhatt
MANAGING EDITOR
Maureen Burke
SENIOR EDITORS
Marjorie Henriquez, Jeff Kearns,
Nicholas Owen, Peter Walker
ASSISTANT EDITORS
Smita Aggarwal, Andrew Stanley
IMF
WEB EDITOR
Rekia Ennaboulssi
PRODUCTION MANAGER
Melinda Weir
PODCASTS
COPY EDITOR
Lucy Morales
SIGN UP TODAY! © 2023 by the International Monetary Fund. All rights reserved.
For permission to reproduce any F&D content, submit a request via
online form (imf.org/external/terms.htm) or by e-mail to copyright@
imf.org. Permission for commercial purposes also available from the
Copyright Clearance Center (copyright.com) for a nominal fee.
Opinions expressed in articles and other materials are those of the
authors; they do not necessarily reflect IMF policy.
Subscriber services, changes of address, and advertising inquiries:
MAJOR PLATFORM
PO Box 92780
Washington, DC 20090, USA
Telephone: (202) 623-7430
APPLE Fax: (202) 623-7201
E-mail: publications@imf.org
GOOGLE Postmaster: send changes of address to Finance & Development, Inter-
national Monetary Fund, PO Box 92780, Washington, DC 20090, USA.
LIBSYN The English edition is printed at Dartmouth Printing Company,
Hanover, NH.
SOUNDCLOUD
Finance & Development is published quarterly by the
SPOTIFY International Monetary Fund,
700 19th Street NW,
Washington, DC 20431,
in English, Arabic, Chinese, French,
Russian, and Spanish.
English edition ISSN 0145-1707
4 D E C E M B E R 2023
F&D
Editor’s Letter
The AI Awakening
full disclosure: this issue of f&d was pro- Policymakers should prepare reforms
duced entirely with human intelligence. But someday for these multiple scenarios and revise
soon at least parts of this magazine may be assisted by as the future unfolds, he notes.
artificial intelligence—a topic that has dominated global This leads us to AI governance. Ian
discourse since ChatGPT’s introduction one year ago. Bremmer, president of Eurasia Group,
Generative AI has introduced tantalizing new possi- and Mustafa Suleyman, CEO of Inflec-
bilities in both the public and private spheres. Think how tion AI, point to regulatory challenges
these “machines of the mind” can improve health care amid a race for AI supremacy among
diagnoses, close education gaps, tackle food insecurity governments. They warn that govern-
with more efficient farming, drive planetary exploration— ing AI will be among the international
not to mention eliminate the drudgery of work. community’s most difficult challenges
Yet the initial excitement surrounding AI has given way in coming decades and outline princi-
to genuine and growing concerns—including about the “Ultimately, ples for AI policymaking.
spread of misinformation that disrupts democracy and
destabilizes economies, threats to jobs across the skills
it’s about what The IMF’s Gita Gopinath urges bal-
ancing innovation and regulation in
spectrum, a widening of the gulf separating the haves and AI can do to developing a unique set of policies for
have-nots, and the proliferation of biases, both human
and computational.
help people.” AI. Because AI operates across borders,
we urgently need global cooperation to
This issue is an early attempt to understand AI’s impli- maximize the enormous opportunities
cations for growth, jobs, inequality, and finance. We bring of this technology while minimizing the
together leading thinkers to explore how to prepare for obvious harms to society, she writes.
an AI world. In other thought-provoking arti-
In our lead article, Stanford’s Erik Brynjolfsson and cles, Daniel Björkegren and Joshua
Gabriel Unger sketch possible “forks in the road” that lead Blumenstock show how Kenya, Sierra
to very different outcomes (beneficial or detrimental) for Leone, and Togo adapted AI to benefit
AI and the economy. The future that emerges will be a the poor. Nandan Nilekani describes
consequence of many things, including technological and how India is on a cusp of an AI revolu-
policy decisions made today, they note. tion to address pressing economic and
For MIT’s Daron Acemoglu and Simon Johnson, AI’s social challenges. And we profile Har-
ultimate impact depends on how it affects workers. Inno- vard labor economist Lawrence F. Katz,
vation always leads to higher productivity, but not always whose defining work on inequality illu-
to shared prosperity, depending on whether machines minates the discussion on AI.
complement or replace humans. The economists out- AI can develop in very different direc-
line policies, such as giving labor a voice, that can redi- tions, underscoring the role of society
rect efforts away from pure automation toward a more in actively and collectively determin-
“human-complementary” path that creates new and high- ing its future. What is clear is that the
er-quality tasks. technology must be guided as tools
AI progresses by leaps and bounds. Given its inherent that can enhance, rather than under-
unpredictability, Anton Korinek, of the University of Vir- mine, human potential and ingenuity.
ginia, recommends scenario planning. He lays out how dif- Ultimately, it’s about what AI can do to
PH OTO: IM F
DECEMBER 2023 5
F&D
T H E BI G P I CTURE: Still emerging from the pandemic, countries in sub-Saharan Africa have been hit by a sluggish
global economy, worldwide inflation, high borrowing costs, and a cost-of-living crisis, says the IMF’s latest Regional
Economic Outlook for Sub-Saharan Africa. Moreover, political instability is an ongoing concern. To ensure that the
rebound is more than just a transitory glimpse of sunlight, the region’s governments must focus on reforms to recover
lost ground from recent crises and create space to address pressing development needs, the report says. Above,
Gladys Lampey at Jamestown Harbor, Accra, Ghana. IMF Photo/Andrew Caballero-Reynolds.
“
Reset in Fiscal the insurer of first resort for all shocks.
They should rebuild depleted fiscal buf-
economies with
the most vulnerable. Revenues must keep
up with spending, and carbon pricing pol-
aging populations,
the temptation to finance all icies should be pursued. Moreover, fiscal
spending through debt must be resisted, frameworks need strengthening; coun-
entitlement reforms
IMF Deputy Managing Director Gita tries must be able to respond to shocks
Gopinath warns in a Financial Times but with clear exit mechanisms.
are inescapable.”
op-ed. She calls for renewed focus on fis- “These are demanding times for pol-
cal policy, and with it, a reset in fiscal pol- icymakers,” Gopinath writes. “Put-
icy thinking. ting fiscal houses in order is essential
We need to rethink what governments to ensure governments can deliver for —Gita Gopinath, The Financial
can do, Gopinath says. They cannot be their people.” Times, October 27, 2023
6 D E C E M B E R 2023
Kaleidoscope F&D
Overheard
4.5 %
RESHORING FRIEND-SHORING
0
WORLD
ASIA
CHINA
WORLD
ASIA
CHINA
-2
“What India needs is a stra- of global output could be lost
tegic plan to chase down the in the long term under a
reshoring scenario, in which
most important opportuni- OECD countries and China
ties for AI to help. The trick -4 increase non-tariff barriers
on all countries to reduce
is not to look too hard at the dependence on foreign inputs
technology but to look at the
problems people face that -6
existing technology has been
unable to solve.”
—Nandan Nilekani, chairman -8
and cofounder of Infosys (see
“Unlocking India’s Potential
with AI,” in this issue of F&D) SOURCE: IMF Regional Economic Outlook for Asia and Pacific, October 2023.
DECEMBER 2023 7
F&D
Back to Basics
Artificial Intelligence’s
this progress, a piece of the puzzle was
still missing. Machines could assist and
predict, but they couldn’t truly under-
poised to unleash a new wave of cre- With innovations such as Google Trans- Economics and finance
ativity and productivity. But it also raises late, digital assistants like Alexa and AI is not, of course, a new concept in
important questions for humanity. Key Siri, and the emergence of self-driving economics and finance. Traditional AI
innovation milestones marked the path cars, machines started to understand (advanced analytics, machine learn-
to its current sophistication. and interact with the world. Yet for all ing, predictive deep learning) has been
8 D E C E M B E R 2023
Back to Basics F&D
crunching numbers, gauging market They say that, like a stochastic parrot, incite panic, and even destabilize eco-
trends, and customizing financial prod- AI can create nonsensical and untrue nomic or financial systems with unprec-
ucts for a long time. What sets GenAI facts, a phenomenon called “halluci- edented efficiency and intensity. It may
apart is its ability to delve deeper and nation,” and it doesn’t really know the not always be deliberate: machines may
interpret complex data in a more cre- meaning behind the words. ChatGPT’s spread misinformation unintentionally
ative manner. By dissecting intricate knowledge, they point out, is limited as a result of hallucinations.
relationships between economic indi- to its latest training date. Possibly. But The threat of AI is not limited to
cators or financial variables, it spits out given the mind-boggling pace of inno- manipulation. Job displacement is
not just forecasts but alternate scenar- vation, how long will these arguments another concern as GenAI continues
ios, insightful charts, and even snippets remain relevant? to advance, potentially automating
of code that could significantly change Still, the initial excitement surround- tasks that were previously performed
how the sector operates. ing GenAI has given way to growing by humans, leading to many job losses
The evolution from traditional to and genuine concerns. Traditional and requiring strategies for employment
generative AI has introduced a new challenges associated with AI, such as and retraining.
era of possibilities into both the pub- the amplification of existing biases in Earlier this year, leading AI experts,
lic and private spheres. Governments training data, or the lack of decision including ChatGPT’s creator, cosigned
are beginning to employ these smarter transparency, have taken on renewed a letter warning that “mitigating the
tools to improve citizen services and urgency. New concerns have also arisen. risk of extinction from AI should be a
overcome workforce shortages. Central global priority alongside other socie-
banks are taking note, seeing in GenAI AI weaponized tal-scale risks such as pandemics and
an enhanced capacity for sifting through One particularly alarming risk is nuclear war.” They were echoing con-
vast amounts of banking data to refine GenAI’s remarkable ability to tell sto- cerns expressed decades earlier by Tur-
economic forecasts and better monitor ries that resonate with individuals’ ing, who warned that “there is a danger
risks, including fraud. preexisting beliefs and viewpoints, that machines will eventually take con-
Investment firms are turning to potentially reinforcing echo cham- trol of our lives.”
GenAI to detect subtle shifts in stock bers and ideological silos. Malicious We stand at a crossroads of tech-
prices and market sentiment, draw- actors can leverage this ability not only nology and ethics. GenAI, with its vast
ing from a larger body of knowledge through the written word: in March promise and profound, existential
to propose more creative options, pav- 2022, an AI-generated video purported questions, cannot be uninvented. As
ing the way for potentially more lucra- to show Ukrainian President Volody- we leverage its transformative power,
tive investment strategies. Meanwhile, myr Zelenskyy surrendering to Rus- it’s imperative to remember Turing’s
insurance companies are exploring how sian forces. Such incidents demon- enduring counsel. GenAI is a monu-
generative models can create personal- strate how GenAI can be weaponized mental shift that demands vigilant over-
ized policies that align more closely with to manipulate politics, markets, and sight, new regulatory frameworks, and
individual needs and preferences. public opinion. an unwavering commitment to ethical,
GenAI is evolving at a breakneck Whether it’s a fabricated story, doc- transparent, controllable innovations
pace, pushing the boundaries of AI tored image, or synthetic video, GenAI that harmonize with human values. F&D
capabilities in economics and finance creations can be so convincing that they
and introducing novel solutions to old create a false sense of reality. This has hervé tourpe is head of the IMF’s
challenges. Some people are skeptical. the potential to spread misinformation, Digital Advisory Unit.
DECEMBER 2023 9
F&D
Point of View
A
R ICH AR D JO PSO N (L EF T ) AN D CO U RT ESY O F M U STA FA SU LEY MAN ( RIG HT )
rtificial intelligence will open people’s lives and when, not if, and a rare instance of
societies to groundbreaking scientific advances, bipartisan consensus.
unprecedented access to technology, toxic mis- This flurry of activity is encouraging.
information that disrupts democracies, and eco- In a remarkably short amount of time,
nomic upheaval. In the process, AI will trigger a funda- world leaders have prioritized the need
mental shift in the structure and balance of global power. for AI governance. But agreeing on
This creates an unparalleled challenge for political the need for regulation is table stakes.
institutions around the world. They will have to establish Determining what kind of regulation is
new norms for a dynamic novel technology, mitigate its just as important. AI doesn’t resemble
potential risks, and balance disparate geopolitical actors’ any previous challenge, and its unique
interests. Increasingly, these actors will come from the
private sector. And it will require a high level of coordina-
AI may become characteristics, coupled with the geo-
political and economic incentives of
tion from governments, including strategic competitors the first the principal actors, call for creativity
and adversaries.
In 2023, governments around the world woke up to this technology in governance regimes.
AI governance is not just one problem.
challenge. From Brussels to Beijing to Bangkok, lawmak-
ers are busy crafting regulatory frameworks to govern AI,
with the means When it comes to climate change, there
may be many routes to achieving the ulti-
even as the technology itself advances exponentially. In
Japan, Group of Seven leaders launched the “Hiroshima
to improve on mate objective of lowering greenhouse
gas emissions, but there is a single over-
Process” to tackle some of the trickiest questions raised by itself. riding objective. AI is different, as an
DECEMBER 2023 11
F&D Point of View
P
governance regime is insufficient eople worry that artificial intelligence is, or will
to address the various sources of soon be, undermining democracy. They fear AI
AI risk. In practice, determining will take away jobs, destabilize the economy, and
which tools are appropriate to target widen the divide between the rich and the poor.
which risks will require developing a This could further concentrate power in the hands of a
live, working taxonomy of discrete few tech companies and weaken government structures
potential AI impacts. AI governance designed to regulate them. Some also fear that tech giants
must therefore be targeted, risk- and government may increasingly delegate human deci-
based, and modular rather than one- sion-making to machines, eventually replacing democ-
size-fits-all. racy with “algocracy,” rule not by the people but by algo-
rithm.
Governing AI will be among the inter- This dystopian vision misses our current capacity to
national community’s most difficult shape AI development. We, as human societies, have the
challenges in the coming decades. As political ability (at least for now) and the responsibility to
important as the imperative to regu- address the harm AI could inflict on us. We also have the
late AI is the imperative to regulate it technological opportunity to harness AI to enhance our
correctly. Current debates on AI pol- democracy in a way that strengthens our collective ability
icy too often tend toward a false debate to govern—rather than simply regulate—AI.
between progress and doom (or geopo- Like other ethical and political challenges, such as
litical and economic advantages versus gene editing, AI governance requires not just more
risk mitigation). And rather than think expert intervention and regulation but more citizen voice
creatively, solutions too often resem- and input—for example, on how to navigate the distribu-
ble paradigms for yesterday’s problems. tive impact of AI on the economy. Like other global con-
This will not work in the age of AI. cerns, such as climate change, AI governance requires
Good policymaking will be vital, but this democratic voice to be heard at the level of interna-
PH OTO COU R TE SY OF HE LE NE L AN DE MOR E
getting there rests on good institutions. tional institutions. Luckily, AI has the potential to usher
To build these institutions, the inter- in a more inclusive, participatory, and deliberative form
national community will need to agree of democracy, including at the global scale.
on a conceptual framework for how to
think about AI. We offer these princi- Participatory experiments
ples as a start. F&D For 40 years many governments have engaged in exper-
iments aiming to include ordinary citizens in policymak-
ian bremmer is president and ing and lawmaking in richer ways than through voting
founder of Eurasia Group and GZERO alone. These experiments have mostly been local and
Media. mustafa suleyman is small-scale, much like the citizens’ assemblies and juries
CEO and cofounder of Inflection AI. that have proliferated on climate and other issues. A 2020
12 D E C E M B E R 2023
Point of View F&D
“AI regulation is
likely to be better
enforced and
more effective in
AI-empowered
democracies.”
Organisation for Economic Co-oper- Politicians, overwhelmed by the raw Over the past few months pol.is has
ation and Development report found and multifaceted data or unsure of its evolved to integrate machine learning
close to 600 such cases in which a ran- meaning, have as a result easily ignored with some of its functions to render
dom sample of citizens engages deeply the citizens’ input. People were allowed the experience of the platform more
with an issue and formulates informed to speak but were not always heard. And deliberative. Contributors to the plat-
policy recommendations (and in one the level of deliberation, even for those form can now engage with a large lan-
case even proposals). involved, was often superficial. guage model, or LLM (a type of AI), that
But some of these political experi- speaks on behalf of different opinion
ments have also aimed for mass partic- Enhanced deliberation clusters and helps individuals figure out
ipation, as in the participatory consti- We now have the chance to scale and the position of their allies, opponents,
tutional processes organized in Brazil, improve such deliberative processes and everyone in between. This makes
Kenya, Nicaragua, South Africa, and exponentially so that citizens’ voices, the experience on the platform more
Uganda in the 1980s and 1990s, and in all their richness and diversity, can truly deliberative and further helps
more recently in Chile, Egypt, and Ice- make a difference. Taiwan Province of depolarization. Today, this tool is fre-
land, which have used mass consulta- China exemplifies this transition. quently used to consult with residents,
tions and crowdsourcing to reach out Following the 2014 Sunflower Rev- engaging 12 million people, or nearly
to ordinary people. Not every attempt olution there, which brought tech- half the population.
has been successful, of course, but all savvy politicians to power, an online Corporations, which face their own
are part of a significant trend. open-source platform called pol.is governance challenges, also see the
Some governments have also rolled was introduced. This platform allows potential of large-scale AI-augmented
out broad multi-format consultation people to express elaborate opinions consultations. After launching its more
campaigns. The 2019 Great National about any topic, from Uber regula- classically technocratic Oversight
Debate launched by French President tion to COVID policies, and vote on Board, staffed with lawyers and experts
Emmanuel Macron in response to the the opinions submitted by others. It to make decisions on content, Meta
yellow vest movement, with some 1.5 also uses these votes to map the opin- (formerly Facebook) began experi-
million participants, is one example. ion landscape, helping contributors menting in 2022 with Meta Commu-
Another is the EU-wide Conference understand which proposals would nity Forums—where randomly selected
on the Future of Europe, which invited garner consensus while clearly iden- groups of users from several countries
citizens from EU member countries to tifying minority and dissenting opin- could deliberate on climate content
weigh in on reforms to EU policies and ions and even groups of lobbyists with regulation. An even more ambitious
institutions, prompting 5 million peo- an obvious party line. This helps peo- effort, in December 2022, involved
ple to visit the website and 700,000 to ple understand each other better and 6,000 users from 32 countries in 19
engage in debate. reduces polarization. Politicians then languages to discuss cyberbullying in
Despite some online elements, these use the resulting information to shape the metaverse over several days. Delib-
have been mostly low-tech, analog pro- public policy responses that take into erations in the Meta experiment were
cesses, involving no AI whatsoever. account all viewpoints. facilitated on a proprietary Stanford
14 D E C E M B E R 2023
F&D
India is on the brink of a transformation that could tion is of working age. This is the start-
ing point: India is a very large country of
change its economic and social future very young people.
This demographic dividend, favor-
able global trends, and the unlocking
of decades of suppressed potential are
starting to show returns. Even as the
macroeconomic projections for most of
the world seem modest or bleak, India
remains a bright spot. These young Indi-
ans are aspirational and motivated to use
every opportunity to better their lives.
What really sets India apart from
the West are its unique challenges and
needs. India’s diverse population and
complex socioeconomic concerns mean
that AI there is not just about develop-
ing cutting-edge technology. It’s about
finding innovative solutions to address
pressing problems in health care, edu-
cation, agriculture, and sustainability.
Though our population is just double
B
efore the end of this decade, more Indians will the size of Europe’s, we are much more
use AI every day than in any other country in the diverse. Indians, like Europeans, are
world. What’s more, people in advanced econo- often bi- or multilingual. India recog-
mies will be surprised by the ways the country will nizes 19,500 dialects spoken by at least
use AI. India is on the cusp of a technological revolution 10,000 people. Based on data from the
that could alter the trajectory of its social and economic Indian census, two Indians selected at
future, and in this revolution there are lessons for the rest random have only a 36 percent chance
of the world. of speaking a common language.
Our prediction hinges on three facts: India needs it, This language barrier is complicated
India is ready for it, and India will do it.
The rest of by the fact that the official literacy rate
in the country hovers near 77 percent,
India needs it
The concept of “China plus one” has been gaining trac-
the world has varying vastly between states. This
means that roughly 1 in 4 people can’t
tion, with its admonition that global companies should
not depend inordinately on China for their manufacturing
been eyeing AI read or write. Even though the gov-
ernment tries to provide welfare assis-
and software needs. India, with its growing infrastructure with curiosity, tance for its most vulnerable, it’s hard to
investments, favorable policies, and young working popu-
lation, is the most likely beneficiary of this shift. It is per- waiting for spread awareness about the service and
reach the last mile. Filling out a simple
haps the only country poised to match the scale of China.
With 1.4 billion people, India is closer to a continent
real-use cases. form to access welfare can be daunting
for someone who is illiterate. Deter-
than a country. Its population is almost twice that of
Europe. But the average age in India is 28, compared with
In India, we see mining eligibility for assistance means
depending on someone who can read,
Europe’s 44, which means a higher share of the popula- potential today. write, and navigate the bureaucracy.
Actually receiving services means assis- about 60 percent of real-time payment at nearly $250 billion a year. Next to
tance seekers must have an agent help- transactions worldwide. those from the US, the largest number
ing them who is not misinformed—or With the success of these models, of developers on GitHub, a cloud-based
worse, corrupt. These barriers dispro- India is embracing innovation in open service for software development, are
portionately affect those who need gov- networks as digital public infrastruc- from India. This sector not only inno-
ernment assistance the most. ture. Take the example of Namma Yatri, vates but also widely adopts digital pub-
We have the ability to solve a lot of a ride-hailing network built in collabo- lic infrastructure. The effect is cyclical:
problems for our population, but the ration with the union of auto-rickshaw start-ups feed the growing tech culture
hard part has always been in the dis- drivers in Bangalore and launched in and, in turn, leverage the data to build
tribution, not the solution. In India, November 2022. These drivers have more precise and beneficial AI tools.
we believe that AI can help bridge this their own app, with a flat fee to use it, no India’s dynamic start-up ecosystem,
access gap. AI enables people to access percentage commission, and no mid- moreover, is actively working on AI
services directly with their voice using dleman. The app has facilitated close to solutions to address various challenges.
natural language, empowering them to 90,000 rides a day, almost as many as AI can be a game changer in edu-
help themselves. As Canadian writer ride-hailing companies in the city. cation as well, helping close the liter-
William Gibson aptly said, “The future Unlike Western countries, which acy gap. AI technologies are uniquely
is already here—it’s just not evenly dis- have legacy systems to overhaul, India’s positioned to help students learn in
tributed.” Nowhere is this more glar- tabula rasa means that AI-first systems their native languages, as well as learn
ingly evident than in India. can be built from the ground up. The English. AI’s applications are useful
The rest of the world has been eye- quick adoption of digital public infra- not only for students; they extend to
ing AI with curiosity, waiting for real- structure is the bedrock for these tech- teachers, who are often overwhelmed
use cases. In India, we see potential nologies. Such infrastructure generates by administrative tasks that detract
today. While this may be true of many enormous amounts of data, and thanks from teaching. As AI takes over routine
other developing economies, the other to India’s Account Aggregator frame- tasks in government and start-ups, the
important factor is that. . . work, the data remain under the citi- roles of teachers and students evolve,
zens’ control, further encouraging pub- and they form dynamic partnerships
India is ready for it lic trust and utilization. With this solid focused on deep learning and mean-
India’s population isn’t just young, it footing, India is well positioned to lead ingful human interaction.
is connected. According to the coun- the charge in AI adoption. What India needs is a strategic plan
try’s telecommunications sector regu- to chase down the most important
lator, India has more than 790 million India will do it opportunities for AI to help. The trick
mobile broadband users. Internet pen- In September 2023, the Indian govern- is not to look too hard at the technol-
etration continues to increase, and with ment, in collaboration with the EkStep ogy but to look at the problems people
the availability of affordable data plans, foundation, launched the PM-Kisan face that existing technology has been
more and more people are online. This chatbot. This AI chatbot works with unable to solve. And organizations such
has created a massive user base for AI PM-Kisan, India’s direct benefit transfer as EkStep have stepped up with a mis-
applications and services. program for farmers, initiated in 2019 sion called People+AI. Instead of put-
But where India has surpassed all to extend financial help to farmers who ting AI first, they focus on the problems
others is in its digital public infrastruc- own their own land. Access to the pro- of people. This has led to surprising new
ture. Today, nearly every Indian has a gram, getting relevant information, and uses unique to India.
digital identity under the Aadhaar sys- resolving grievances was always a prob- India’s emerging status as a tech-
tem. The Aadhaar is a 12-digit unique lem for the farmers. The new chatbot nological powerhouse, combined with
identity number with an option for gives farmers the ability to know their its unique socioeconomic landscape,
users to authenticate themselves digi- eligibility and the status of their applica- puts it in a favorable position to be the
tally—that is, to prove they are who they tion and payments using just their voice. world’s most extensive user of AI by the
claim to be. On launch day more than 500,000 users end of this decade. From streamlining
Further, India set up a low-cost, chatted with the bot, and features are education to aiding in social protec-
real-time, interoperable payment sys- being released slowly to ensure a safe tion programs, AI has the potential to
tem. This means that any user of any and risk-managed rollout. deeply penetrate Indian society, effect-
bank can pay any other person or mer- These steps are part of an encourag- ing broad and meaningful change. F&D
chant using any other bank instantly ing trend of early adoption of new tech-
and at no cost. This system—the Uni- nology by the Indian government. But nandan nilekani is the chairman
fied Payments Interface—handles the trend extends beyond the govern- and cofounder of Infosys and founding
more than 10 billion transactions a ment. India’s vibrant tech ecosystem chairman of UIDAI (Aadhaar). tanuj
month. It is the largest real-time pay- has taken off as well, a direct offshoot bhojwani is head of People+AI.
ment system in the world and handles of its booming IT exports—currently
16 D E C E M B E R 2023
IMF WEEKEND READ
SIGN UP TODAY!
F&D Picture This
their impact.
97%
This safety net consists of four main
layers: countries’ own international
reserves, bilateral swap lines whereby
central banks exchange currencies to
provide liquidity to financial markets, of international reserves
regional financing arrangements by are held by approximately
which countries pool resources to lever- half the world’s economies,
with a group of about
age financing in a crisis, and the IMF. 90 vulnerable emerging
International reserves are the first market and low-income
line of defense in a crisis; however, countries accounting for
the remaining 3%.
because of their high cost, they are
unevenly distributed, with most held by SOURCE: IMF.
andrew stanley is on the staff of SOURCES: Central bank websites; regional financing arrangements' annual reports; and IMF staff estimates.
Finance & Development. NOTE: Since the safety net is composed of various currencies, its US dollar value fluctuates with exchange rates.
18 D E C E M B E R 2023
Picture This F&D
Shrinking safeguard
Although international reserves have risen Distribution of international reserves
rapidly, that increase in self-insurance has (US dollars, latest available data)
been highly uneven. Poorer countries remain
underinsured, leaving them vulnerable to
shocks. Meanwhile, the IMF, at the center
of the safety net, has shrunk in relation
to the total size of global external
liabilities and is now far more reliant on
resources temporarily borrowed from a few
member states. The IMF’s traditional quota
resources, its permanent capital contributed
by all member states, have decreased in
relative terms.
International reserves
(trillions of US dollars)
SOURCES: Central bank websites; regional financing arrangements' annual reports; and IMF staff calculations.
NOTE: Reserves chart and map above exclude gold. Data are not available for all economies. The boundaries, colors, denominations, and any other information shown on the
maps do not imply, on the part of the IMF, any judgment on the legal status of any territory or any endorsement or acceptance of such boundaries.
DECEMBER 2023 19
F&D Artificial Intelligence
THE MACROECONOMICS OF
ARTIFICIAL INTELLIGENCE
The collective decisions we make
today will determine how AI affects
productivity growth, income inequality,
and industrial concentration
DECEMBER 2023 21
F&D Artificial Intelligence
Low-productivity future
On one path of the productivity fork, AI’s impact is lim-
ited. Despite the rapidly improving technical capabili-
E
ties of AI, its adoption by businesses may continue to
be slow and confined to large firms (Zolas and others
2021). The economics of AI may turn out to be of a very
narrow labor-saving variety (what Daron Acemoglu
conomists have a poor track record of predict- and Simon Johnson call a “so-so technology,” such as
ing the future. And Silicon Valley repeatedly an automated grocery checkout stand), instead of one
cycles through hope and disappointment over that enables workers to do something novel or powerful
the next big technology. So a healthy skepti- (see “Rebalancing AI” in this issue of F&D). Displaced
cism toward any pronouncements about how workers might disproportionately end up in even less
artificial intelligence will change the economy is jus- productive and less dynamic jobs, further muting any
tified. Nonetheless, there are good reasons to take aggregate benefit to the long-term productivity growth
seriously the growing potential of AI—systems that rate of the economy.
exhibit intelligent behavior, such as learning, reason- Like so many of Silicon Valley’s recent technologi-
ing, and problem-solving—to transform the economy, cal enthusiasms (3D printers, self-driving cars, virtual
especially given the astonishing technical advances reality), AI may also end up being less promising or less
of the past year. ready to bring to market than initially hoped. Any real
AI may affect society in a number of areas besides economic gains, even modest ones, may show up in
the economy—including national security, politics, and the data many decades after the first moments of tech-
culture. But in this article, we focus on the implications nological promise, as has often been the pattern. The
of AI on three broad areas of macroeconomic interest: famous paradox identified by economist Robert Solow
productivity growth, the labor market, and industrial in 1987—“You can see the computer age everywhere
concentration. AI does not have a predetermined future. but the productivity statistics”—may become more
It can develop in very different directions. The particu- extreme, as everyone seems to have an AI chatbot that
lar future that emerges will be a consequence of many amazes their friends, but businesses do not seem more
things, including technological and policy decisions productive for their increased use of AI. Firms may fur-
made today. For each area, we present a fork in the road: ther blunt any economic benefits from AI by failing to
two paths that lead to very different futures for AI and figure out the organizational and managerial changes
the economy. In each case, the bad future is the path of they need to best leverage it.
least resistance. Getting to the better future will require And, as in the case of self-driving cars, the techno-
good policy—including logical challenges of going from an exciting proof of
• Creative policy experiments concept to a highly reliable product may be further
• A set of positive goals for what society wants from AI, compounded by a legal regime that was not designed
not just negative outcomes to be avoided to accommodate this new technology and may seri-
• Understanding that the technological possibilities ously hinder its development. In the case of AI, there
of AI are deeply uncertain and rapidly evolving and is tremendous uncertainty over what current laws con-
that society must be flexible in evolving with them cerning intellectual property imply when models are
trained on millions of data points that may include the
First fork: Productivity growth
The first road concerns the future of economic growth—
which is largely the future of productivity growth. The
US economy has been stuck with disturbingly low pro-
industrial concentration.”
productivity growth may be the globe’s most funda-
mental economic challenge.
22 D E C E M B E R 2023
Artificial Intelligence F&D
creativity and scientific discovery itself—math, sci-
ence, further AI development—a kind of recursive
self-improvement that was once just a science fiction
thought experiment.
DECEMBER 2023 23
F&D Artificial Intelligence
24 D E C E M B E R 2023
Artificial Intelligence F&D
the residual control rights to their work might be one science labs at universities will also develop AI models,
in which more innovators decide they would rather be some of which they will make open-source. Federal leg-
owners of small firms than be employees of large ones. islators and regulators will have a large impact, as might
The result is that the long rise in industrial concen- more local ones. Voters have a voice. Labor unions must
tration starts to run aground, because some nimble figure out what kind of relationship they want with AI
smaller businesses close or even reverse the technology and what their demands will be.
gap with their larger counterparts and win back more Although we have sketched a number of possible
market share. futures for AI, we want to emphasize not only how
deeply unpredictable the future of this technology is
Toward a policy agenda but also the agency society has in actively and collec-
For each of the forks in the road, the path that leads to tively determining which AI future emerges.
a worse future is the one of least resistance and results We have raised more questions than we have
in low productivity growth, higher income inequality, answered, which reflects, in part, the nascent stage of AI
and higher industrial concentration. Getting to the good adoption and impact. But it also reflects a deeper imbal-
path of the fork will require hard work—smart policy ance between research efforts advancing the frontier of
interventions that help shape the future of technology the technology and the more limited research aimed at
and the economy. understanding its economic and social consequences.
It is also important to appreciate a broader point This imbalance was of less significance when the
about policy. Much of the discourse around AI regula- technology had limited macroeconomic consequences.
tion now takes place along a kind of hydraulic model: But today, when the effects of AI on society are likely
should we have more AI or less AI—or even ban AI. This to be measured in trillions of dollars, far greater invest-
discussion happens when AI is perceived as somewhat ment should be made in research on the economics of
of a fixed thing, with a predetermined future. AI can AI. Society needs innovations in economic and policy
come fast or slow. There can be more or less of it, but understanding that match the scale and scope of the
basically it is what it is. breakthroughs in AI itself. Reorienting research priori-
However, if policymakers understand that AI can ties and developing a smart policy agenda can help soci-
develop in different directions, the discourse will be ety move toward a future of both sustained and inclusive
framed differently. How can policies encourage the economic growth. F&D
types of AI that complement human labor instead of
imitating and replacing it? What choices will encourage erik brynjolfsson is the Jerry Yang and Akiko
the development of AI that firms of all sizes can access, Yamazaki Professor at the Stanford Institute for Human-
instead of just the largest ones? What kind of open- Centered AI, where he directs the Stanford Digital
source ecosystem might that require, and how do policy- Economy Lab. gabriel unger is a postdoctoral
makers support it? How should AI labs approach model fellow at the Stanford Digital Economy Lab.
development, and how should firms approach AI imple-
mentation? How does society get an AI that unleashes references
radical innovation, instead of marginal tweaks to exist- Autor, David, Frank Levy, and Richard Murnane. 2003. “The
ing goods, services, and systems? Skill Content of Recent Technological Change.” Quarterly
Journal of Economics 118 (4): 1279–333.
Many different actors have power to affect the direc-
Brynjolfsson, Erik, Daniel Rock, and Chad Syverson. 2019.
tion of the AI future. Major corporations will have to
“Artificial Intelligence and the Modern Productivity Paradox:
make important decisions about how they choose to A Clash of Expectations and Statistics.” In The Economics of
integrate AI into their workforce. The largest of these Artificial Intelligence: An Agenda, edited by Ajay Agrawal, Joshua
companies will also develop in-house AI. AI/computer Gans, and Avi Goldfarb. Chicago: University of Chicago Press.
Brynjolfsson, Erik, Danielle Li, and Lindsay Raymond.
2023. “Generative AI at Work.” NBER Working Paper 31161,
National Bureau of Economic Research, Cambridge, MA.
“Society needs innovations Eloundou, Tyna, Sam Manning, Panels Mishkin, and Daniel
Rock. 2023. “GPTs Are GPTs: An Early Look at the Labor Mar-
the scale and scope of the Technologies Adoption and Use by U.S. Firms: Evidence from
the Annual Business Survey.” NBER Working Paper 28290,
breakthroughs in AI itself.”
National Bureau of Economic Research, Cambridge, MA.
https://www.nber.org/papers/w28290.
DECEMBER 2023 25
F&D Artificial Intelligence
REBALANCING
AI
Daron Acemoglu and Simon Johnson
26 D E C E M B E R 2023
Artificial Intelligence F&D
let AI happen,” which focuses primarily on auto-
mation (replacing people). We must recognize
that there is no singular, inevitable path of devel-
opment for new technology. And, assuming that
the goal is to sustainably improve economic out-
comes for more people, what policies would put AI
development on the right path, with greater focus
on enhancing what all workers can do?
New functions
The creation of new tasks is even more important
for raising worker marginal productivity. When
new machines open up new uses for human labor,
this expands workers’ contributions to production
F IRST NA ME L AST
DECEMBER 2023 27
F&D Artificial Intelligence
duction methods and assembly lines simultane- small improvements in productivity, but its main
ously introduced a range of new design, technical, function is to extract more effort from workers.
machine-operation, and clerical tasks, boosting the All this underscores perhaps the most import-
industry’s demand for workers. ant aspect of technology: choice. There are often
New tasks have been vital in the growth of myriad ways of using our collective knowledge to
employment and wages over the past two centuries. improve production and even more ways to direct
And many of the fastest-expanding occupations in innovation. Will we invent and implement digital
the past few decades—those of MRI radiologists, tools for surveillance, automation, or to empower
network engineers, computer-assisted machine workers by creating new productive tasks?
operators, software programmers, IT security per- When the productivity bandwagon is weak and
sonnel, and data analysts—did not exist 80 years there are no self-correcting mechanisms to ensure
ago. Even people in occupations that have been shared benefits, these choices become more conse-
around longer, such as bank tellers, professors, quential—and a few tech decision-makers become
and accountants, now work on many relatively economically and politically more powerful.
new tasks using technology. In almost all these
cases, new tasks were introduced because of tech- Complementing humans
nological advances and have been a major driver New technology may complement workers by
of employment growth. These new tasks have also enabling them to work more efficiently, perform
been integral to productivity growth—they have higher-quality work, or accomplish new tasks. For
helped launch new products and enabled more example, even as mechanization gradually pushed
efficient production processes. more than half of the US labor force out of agricul-
ture, a range of new blue-collar and cleri-
Productive automation “AI offers an cal tasks in factories and newly emerging
Automation in an industry can also drive up employ-
ment—in that sector or in the economy broadly—if opportunity to service industries generated significant
demand for skilled labor between about
it substantially increases productivity. In this case,
new jobs may come either from nonautomated
complement worker 1870 and 1970. This work was not only
better paying but also less dangerous
tasks in the same industry or from the expansion
of activities in related industries. In the first half of
skill and expertise and less physically exhausting.
This virtuous combination—auto-
the 20th century, the rapid increase in car manu- if we direct mation of traditional work alongside
facturing stimulated massive expansion of the oil,
steel, and chemical industries. Vehicle output on its development creation of new tasks—proceeded in rel-
ative balance for much of the 20th cen-
a mass scale also revolutionized the possibilities
for transportation, enabling the rise of new retail,
accordingly. ” tury. But sometime after approximately
1970, this balance was lost. While auto-
entertainment, and service activities. mation has maintained its pace or even accelerated
The productivity bandwagon is not activated, over the ensuing five decades, the offsetting force of
however, when the productivity gains from auto- new task creation has slowed, particularly for work-
mation are small—what we call “so-so automa- ers without four-year college degrees. As a result,
tion.” For example, self-checkout kiosks in gro- these workers are increasingly found in low-paying
cery stores bring limited productivity benefits (though socially valuable) services such as in clean-
because they merely shift the work of scanning ing, food service, and recreation.
items from employees to customers. When stores The critical question of the new era of AI is
introduce self-checkout kiosks, fewer cashiers whether this technology will primarily accelerate
are employed, but there is no major productivity the existing trend of automation without the off-
boost to stimulate the creation of new jobs else- setting force of good job creation—particularly for
where. Groceries do not become much cheaper, non-college-educated workers—or whether it will
there is no expansion in food production, and instead enable the introduction of new labor-com-
shoppers do not live differently. plementary tasks for workers with diverse skill sets
Even nontrivial productivity gains from automa- and a wide range of educational backgrounds.
tion can be offset when they are not accompanied by It is inevitable that AI systems will be used for
new tasks. For example, in the American Midwest, some automation. A major barrier to automation
the rapid adoption of robots has contributed to mass of many service and production tasks has been that
layoffs and ultimately prolonged regional decline. they require flexibility, judgment, and common
The situation is similarly troubling for work- sense—which are notably absent from pre-AI forms
ers when new technologies focus on surveillance. of automation. Artificial intelligence, especially
Increased monitoring of workers may lead to some generative AI, can potentially master such tasks.
28 D E C E M B E R 2023
Artificial Intelligence F&D
It is unclear how much this type of automation will • Labor voice: Given that workers will be pro-
contribute to aggregate productivity growth while foundly affected by AI, they should have a voice
these technologies are immature, but they could in its development. Government policy should
contribute to sizable productivity gains as costs fall restrict deployment of untested (or insufficiently
and reliability improves. tested) AI for applications that could put workers
The dominant intellectual paradigm in today’s at risk, for example in high-stakes personnel deci-
digital tech sector also favors the automation path. sion-making tasks (including hiring and termina-
A major focus of AI research is to attain human tion) or in workplace monitoring and surveillance.
parity in a vast range of cognitive tasks and, more • Funding for more human-complemen-
generally, to achieve artificial general intelligence tary research: Research and development in
that mimics and surpasses human capabilities. This human-complementary AI technologies require
intellectual focus encourages automation rather greater support. Governments should foster com-
than the development of human-complementary petition and investment in technology that pairs
technologies. AI tools with human expertise to improve work
However, AI offers an opportunity to comple- in vital social sectors. Once there is sufficient
ment worker skill and expertise if we direct its progress, governments can encourage further
development accordingly. investment with advice on whether purported
Human productivity is often hampered by lack human-complementary technology is appropri-
of specific knowledge or expertise, which could ate for adoption in publicly funded education and
be supplemented by next-generation technology. health care programs.
For example, AI holds great potential for training • AI expertise within government: AI will touch
and retraining expert workers, such as educators, every area of government investment, regula-
medical personnel, and those in modern crafts tion, and oversight. Developing a consultative
(such as electricians and plumbers). AI could also AI division within government can help agencies
create new demands for human expertise and and regulators support more timely, effective
judgment in overseeing these processes, com- decision-making.
municating with customers, and enabling more
sophisticated services. Potential macroeconomic impact
AI could increase global GDP over the next five
Five principles years, although not as substantially as enthusiasts
Redirecting technological change is not easy, but it claim. It might even modestly raise GDP growth in
is possible. Governments everywhere—especially the medium term. However, on our current trajec-
in the US and other countries where technology is tory, the first-order impact is likely to be increased
under active development—should take the fol- inequality within industrial countries.
lowing five steps to help put AI development onto Middle-income countries and many lower-in-
a human-complementary, rather than human-dis- come countries also have much to fear from the
placing, path: existing path. New capital-intensive technology
• Reform business models: The dominant develop- will soon be applied everywhere. There is no guar-
ers of AI easily expropriate consumer data with- antee that, on its current path, AI will generate
out compensation, and their reliance on digital more jobs than it destroys.
advertising incentivizes grabbing consumers’ If we can redirect AI onto a more human-com-
attention through any means possible. Govern- plementary path, while using it to address pressing
ments need to establish clear ownership rights for social problems, all parts of the planet can benefit.
all consumers over their data and should tax digi- But if the just-automate approach prevails, shared
tal ads. Enabling a more diverse range of business prosperity will be even harder to achieve. F&D
models—or even requiring more competition—is
essential if AI is to be helpful to all humans. daron acemoglu is Institute Professor at
• Tax system: The tax code in the US and many the Massachusetts Institute of Technology (MIT).
other countries places a heavier burden on firms simon johnson is the Ronald A. Kurtz
that hire labor than on those that invest in algo- Professor of Entrepreneurship at MIT Sloan and a
rithms to automate work. To shift incentives former IMF chief economist.
toward human-complementary technological
choices, policymakers should aim to create a This article is adapted from the authors’ book,
more symmetric tax structure, equalizing mar- Power and Progress: Our 1000 Year Struggle over
ginal tax rates for hiring (and training) labor and Technology and Prosperity, and also draws on joint
for investing in equipment and software. work with David Autor.
DECEMBER 2023 29
F&D Artificial Intelligence
SCENARIO PLANNING
FOR AN A(G)I FUTURE
Anton Korinek
A RU N SANK A R/ AF P V IA G ET T Y I MAG ES
30 D E C E M B E R 2023
Artificial Intelligence F&D
DECEMBER 2023 31
F&D Artificial Intelligence
32 D E C E M B E R 2023
Artificial Intelligence F&D
mainstream macroeconomic model of automation. range of labor tasks. Levels of investment, such
The results for all three scenarios are illustrated in as investment in research and development, tal-
Chart 2, in which the path of output for each sce- ent, and computer chips, capture how much of our
nario is displayed in panel 1 and the path of com- resources are flowing into AI development. Indica-
petitive market wages in panel 2. tors of growing AI adoption through all sectors of
Three main insights stand out: the economy would capture whether the resulting
First, whereas growth continues along the trajec- systems are usefully deployed in practice. Finally,
tory we are used to from past decades in the conser- the macroeconomic implications would eventually
vative business-as-usual scenario, output growth in become visible in productivity statistics and labor
the two AGI scenarios is much faster, as the scarcity market trends.
of labor is no longer a constraint on output. Tracking these complementary signals allows
Second, wages initially rise in all three scenar- policymakers to tailor policy responses to the real-
ios—but only as long as labor is scarce. They plum- ities of AI as they manifest. But we must remain
met as the economy is close to reaching AGI. humble—the future is likely to surprise us.
Third, the takeoff in output and the collapse in The starkly different economic trajectories
wages in the two AGI scenarios are both driven by implied by the three scenarios described earlier
the same force: the substitution of scarce labor underscore the importance of developing adap-
by comparatively more abundant machines. This tive policy frameworks that can respond nimbly
suggests that it should be possible to design insti- as the future unfolds. Policymakers should stress-
tutions that compensate workers for their income test existing institutions against each scenario
losses and ensure that the gains from AGI lead to and reform them where necessary to ensure they
shared prosperity. are resilient. This may involve gradual steps, such
Chart 2 illustrates the broad con- “Given the starkly as reforming systems of taxation and expanding
tours of how unprecedented tech-
nological changes may affect the differing perspectives social safety nets, or new programs, such as intro-
ducing small basic incomes that can be scaled up
macroeconomy, but it is best under-
stood as an illustration of possibili-
on future progress in when necessary.
Policymakers should charge teams of experts
ties rather than as a precise predic-
tion. A long list of caveats applies.
AI, it would be unwise with iterative scenario planning to help them reg-
ularly update their views on how the probabilities
First, the model underlying the to put all eggs in one of the various scenarios evolve. Embracing the
chart is cast in an efficient econ-
omy in which labor earns competi- basket and formulate uncertainty through an adaptable, scenario-based
approach will allow us to maximize the benefits and
tive returns. A range of factors may
slow the rollout of AGI compared
economic plans for a mitigate the risks in the economic sphere from AI’s
continuing evolution. F&D
with what is technologically possi-
ble, from organizational frictions,
single scenario.” anton korinek is a professor in the
regulations, and constraints on capital accumu- Department of Economics and Darden School
lation—such as chip supply chain bottlenecks—to of Business at the University of Virginia. He is
societal choices on the implementation of AGI. economics of AI lead at the Centre for the Governance
Even when it is technologically possible to replace of AI, a research associate at the National Bureau
workers, society may choose to keep humans in of Economic Research, and a research fellow at the
certain functions—for example, as priests, judges, Centre for Economic Policy Research.
or lawmakers. The resulting “nostalgic” jobs
could sustain demand for human labor in perpe- references
tuity (Korinek and Juelfs, forthcoming). Korinek, Anton.Korinek, Anton. 2023. Generative AI
for Economic Research: Use Cases and Implications for
To determine which AI scenario the future most
Economists. Journal of Economic Literature 2023, 61(4).
resembles as events unfold, policymakers should
Korinek, Anton, and Megan Juelfs. Forthcoming. “Pre-
monitor leading indicators across multiple domains,
paring for the (Non-existent?) Future of Work.” Oxford
keeping in mind that all efforts to predict the pace Handbook of AI Governance. Oxford, UK: Oxford Univer-
of progress face tremendous uncertainty. Useful sity Press.
indicators span technological benchmarks, levels of Korinek, Anton, and Donghyun Suh. 2023. “Scenarios
investment flowing into AI development, adoption for the Transition to AGI.” University of Virginia working
of AI technologies throughout the economy, and paper, Charlottesville, VA.
resulting macroeconomic and labor market trends. Susskind, Daniel. Forthcoming. “Technological Unem-
Technological benchmarks offer the most direct ployment.” Oxford Handbook of AI Governance. Oxford,
measure of how well AI systems perform a wide UK: Oxford University Press.
DECEMBER 2023 33
F&D Artificial Intelligence
TECHNOLOGY’S
BIFURCATED BITE
Andrew Berg, Chris Papageorgiou, and Maryam Vaziri
T echnological developments—such as
factory robots, smart home devices, and
self-driving cars—transform the way we
live and work. Such developments are
exciting in many ways, because they prom-
ise higher productivity and standards of living. But
they can also be frightening: when the machines
take over, how will many people make a living?
This is an old question, of course. Fears about
In retrospect, Kennedy’s concern about lost
jobs seems misplaced. In the years after his speech,
the US economy created millions of net new jobs,
and mass technological unemployment did not
emerge—as demonstrated by today’s unemploy-
ment rate of about 3.5 percent and the multi-de-
cade-high ratio of employment to population.
These labor market developments would seem
to assuage the concerns of a modern-day Luddite:
technology destroying jobs, displacing workers, with the benefits of technology and the power of the
and damaging lifestyles arose during the Industrial market, people will find new jobs, and rising produc-
Revolution—best exemplified, perhaps, by the Lud- tivity will raise living standards—which ultimately
dites in England, who fought life-altering changes happened during the Industrial Revolution of the
in the textile industry. These fears persist today. As 18th and 19th centuries. Indeed, the standard of liv-
then-US Senator John F. Kennedy said in 1960, at ing has increased enormously since 1900. Technolo-
the dawn of the computer revolution, “Today we gies such as electricity, internal combustion engines,
stand on the threshold of a new industrial revolu- telephones, and modern medicine have improved
tion—the revolution of automation. This is a revo- the quality of life and increased life expectancy.
lution bright with the hope of a new prosperity for That is not to say, however, that Kennedy’s con-
labor and a new abundance for America, but it is cerns were unfounded. Only a few years after his
also a revolution which carries the dark menace of speech, wage inequality began to worsen sharply
industrial dislocation, increasing unemployment, (see Chart 1), and the share of income going to
and deepening poverty.” workers fell.
34 D E C E M B E R 2023
Slug F&D
Economists have developed frameworks for A person holds mium, whereas persistent increases in demand for
thinking about the implications of artificial intel- a sign during a such workers had the opposite effect. These forces
rally in support
ligence (AI)—which simulates human intelligence of striking explain both the dip in the skill premium in the early
in machines—and, more generally, the impact of warehouse workers 1970s—when the supply of educated workers rose
technological change, automation, and robots on in Coventry, sharply because more people went to college—and
England, January
inequality. In this respect we will highlight four key 2023. the rise in skill premiums after the 1980s.
channels that affect inequality: In the second channel, capital, especially
• Technological change that improves the pro- machinery and equipment, tends to complement
ductivity of skilled more than unskilled workers skilled workers and substitute for unskilled work-
• Reductions in the cost of capital that comple- ers—for example, machine tools require more
ment chiefly skilled labor programmers but replace other workers in fac-
• Increased ability of machines to replace workers tories. Berg, Buffie, and Zanna (2018) extend this
entirely for particular tasks approach to look at AI and robots as a new type of
• Increased concentration of market power in a capital—additional to traditional machinery and
few firms as a result of technology structures—that substitutes for some groups of
workers and complements others. Over the past
HE NRY NI CHO LLS /R EU T ER S
Regarding the first channel, Katz and Murphy 30 years the substitutability between informa-
(1992) explained the evolution of relative wages in tion and communications technologies (ICT)—a
the United States as the outcome of a race between proxy for new technologies, including computers
increases in the demand and supply of skilled work- and early AI—and unskilled workers seems to have
ers. They focused on aggregate productivity and fac- increased (see Chart 2). In other words, ICT capital
tor-augmenting technological change. Increases in apparently is now better able to perform the tasks
the supply of skilled workers reduced the skill pre- of unskilled workers.
DECEMBER 2023 35
F&D Artificial Intelligence
36 D E C E M B E R 2023
Artificial Intelligence F&D
short run could be a lifetime for some workers (Berg, be reversed as lower-skilled workers benefit more?
Buffie, and Zanna 2018). Or will major corporations—with the best access to
The First Industrial Revolution reflected both data, computers, and top talent—gain more eco-
the optimistic long-term and worrisome short- nomic and political power? The so-far-hypothetical
term perspectives. Few would want to give up the prospect of artificial general intelligence (AGI) adds
benefits from earlier industrial revolutions—from another dose of uncertainty. AGI would presumably
indoor toilets to cell phones—but the transition be capable of any human intellectual effort. How
was both economically and politically wrenching. all this would play out will clearly depend on both
Carl Benedikt Frey argues in The Technology Trap the evolution of the technology and the policy and
that for certain “vulnerable” groups three entire the broader societal response. There are optimistic
generations were worse off as a result. Joseph Sti- and pessimistic AI scenarios, but under any of them,
glitz argues in the December 6, 2011, issue of Van- economic, social, and political upheaval seems a
ity Fair that the technology-driven transition from safe prediction, and policymakers must do their best
agriculture to manufacturing in the 1920s set the to understand the distributional implications of the
stage for the Great Depression. More recently, the rapid changes that are underway.
distributional implications of technological change As we navigate the transition to widespread use
are arguably an important factor in the rise of pop- of AI, it is crucial to acknowledge the global impli-
ulism and anti-globalization sentiment. cations of AI technologies—which so far have not
DATA
AI is rapidly evolving in unpredicted directions— been largely studied. Previous research suggested
perhaps making it impossible to draw any historical
lessons. The early 2023 emergence of ChatGPT-4—
an AI model that seeks to generate human-like
language—marks a significant acceleration in the
25%
Computers may
that the substitution of AI for unskilled labor could
widen global income disparities, putting lower-in-
come countries at a disadvantage (Alonso and oth-
ers 2022). But the advent of generative AI suggests
pace of change, highlighting AI’s ability to extend be able to that the impact of these technologies on different
write at the
far beyond routine tasks. Experts in AI surveyed by level of the countries is uncertain. Developing economies
McKinsey in 2019 expected computers to be able top 25 percent may benefit from AI as a tireless universal tutor
to write at the level of the top 25 percent of humans of humans by and expert programming assistant that strength-
2024, according
by 2050 and perform human-level creative tasks by to AI experts ens their workforces. Conversely, limited access
2055. However, they have revised their estimates to surveyed by to data and expertise and technological gaps could
2024 and 2028, respectively. McKinsey. widen divergence. F&D
It is easy to see why the projections have changed
so sharply. Generative pretrained transformers andrew berg is deputy director of the IMF’s
(GPT) seem to have the potential for widespread Institute for Capacity Development, where
labor market impact—one estimate suggests that maryam vaziri is an economist. chris
once GPT is introduced into the work environment, papageorgiou is a division chief in the IMF’s
about 20 percent of workers could see at least half of Research Department.
their tasks affected. GPT seems to increase produc-
tivity in more creative tasks, such as writing, legal references
analysis, and programming. These studies compare Acemoglu, D., and P. Restrepo. 2020. “Robots and Jobs:
Evidence from US Labor Markets.” Journal of Political
the productivity of groups using GPT with a control
Economy 128 (6): 2188–244.
group in the given task and find big jumps in pro-
Acemoglu, D., and P. Restrepo. 2022. “Demographics and
ductivity with GPT. Just as remarkable, though, is
Automation.” Review of Economic Studies 89 (1): 1–44.
the observation that the least-skilled participants
Autor, D. H. 2019. “Work of the Past, Work of the Future.”
benefit most and that at least in some cases the
In AEA Papers and Proceedings, vol. 109, 1–32. Nashville,
GPT-augmented input is more creative; more- TN: American Economic Association.
over, there are signs that GPT-4 alone may exceed
Berg, Andrew, Edward F. Buffie, Mariarosaria Comunale,
human-level output. These findings contrast with Chris Papageorgiou, and Luis-Felipe Zanna. Forthcom-
earlier emphasis on the automation of routine tasks ing. “Searching for Wage Growth: Policy Responses to
and the substitution of AI and robots for unskilled the AI Revolution.” IMF Working Paper, International
labor. Such shifts in the impact of new technolo- Monetary Fund, Washington, DC.
gies on skilled and low-skilled workers seem to be Berg, A., E. F. Buffie, and L. F. Zanna. 2018. “Should
a key difference between GPT and previous waves We Fear the Robot Revolution? (The Correct Answer Is
of technology, such as digitalization. Yes).” Journal of Monetary Economics 97: 117–48.
All this suggests major implications for both Katz, L. F., and K. M. Murphy. 1992. “Changes in Relative
Wages, 1963–1987: Supply and Demand Factors.” Quar-
growth and inequality, but it also suggests that the
terly Journal of Economics 107 (1) 35–78.
past may not be prologue. Will some wage inequality
DECEMBER 2023 37
F&D Artificial Intelligence
AI’S REVERBERATIONS
ACROSS FINANCE
Jeff Kearns
A
rtificial intelligence tools and the people to use
them are the new must-haves for the world’s
financial institutions and central banks.
In June 2023, JPMorgan Chase & Co. had
3,600 AI help-wanted postings, according to
Evident Insights Ltd., a London-based start-up tracking
AI capabilities across financial services companies.
“There’s a war for talent,” said Alexandra Mousaviza-
deh, the founder of Evident Insights. “Making sure you
are ahead of it now is really life and death.”
M AT T C HIN WO RT H
38 D E C E M B E R 2023
F IRST NA ME L AST
Artificial Intelligence F&D
DECEMBER 2023 39
F&D Artificial Intelligence
assets and predict markets. Or the sector may have For other uses, such as making institutional
the most to lose if AI spurs theft, fraud, cybercrime, decisions on investment and trading, AI can
or even a financial crisis that investors can’t con- be limited by data that proves unreliable or by
ceive of today. unprecedented high-impact situations, she said.
The debut of OpenAI’s ChatGPT in November It’s also a priority to avoid abuses and ensure that
2022 is rippling through finance and other indus- AI is used within a secure, ethical, and compliant
tries. It quickly topped 100 million users to become framework.
the fastest-growing application in internet history. “Artificial intelligence cannot replace the brain,”
In finance, the demand for people who know Defend said. A wholly AI-driven process could be
how to tap into AI is global. Three of the top 10 cities dangerous, she said. “It’s equally important the
in Evident’s talent index are in India, said Mousav- interpretation, the understanding, and the check
izadeh, an economist, mathematician, and former of what the algorithms are providing.”
cohead of country risk at Morgan Stanley. JPMorgan, the largest US lender, spends more
than $15 billion a year on technology, where
Financial embrace it deploys almost a fifth of its approximately
The money flowing into AI from financial and other 300,000 employees. An AI research group
enterprises underscores the new priorities. Sales employs 200, and AI enables hundreds of uses
of software, hardware, and services for AI systems ranging from prospecting and marketing to risk
will climb 29 percent this year to $166 billion and top management and fraud prevention. AI also runs
$400 billion in 2027, according to International Data across payment processing and money movement
Corp. Financial sector spending will more systems worldwide.
than double to $97 billion in 2027, with “AI tools may “It is an absolute necessity,” Chief Executive
a 29 percent compound annual growth
rate—the fastest of five major industries— exacerbate a crisis, Officer Jamie Dimon told shareholders in April.
40 D E C E M B E R 2023
Artificial Intelligence F&D
tion Hub’s Project Aurora, for example, showed that “The technology creeps up on us when we start
neural networks, a type of machine learning, can trusting it but using it more and more,” Daniels-
help detect money laundering by sniffing out pat- son said.
terns and anomalies in transactions that traditional AI could spark a financial crisis, according to US
methods can’t identify. Securities and Exchange Commission Chair Gary
Gensler. He is charged with protecting a $46 trillion
Signal in noise stock market that makes up two-fifths of the world
The Bank of Canada built a machine learning total. Financial stability risks from AI demand “new
tool to detect anomalies in regulatory submis- thinking on systemwide or macroprudential policy
sions. Data Science Director Maryam Haghighi interventions,” he told reporters in July. “AI may
said its automated daily runs catch things people heighten financial fragility, as it could promote
wouldn’t—while freeing up staff to follow up on herding—with individual actors making similar
the analysis. decisions because they are getting the same signal
“This is an example of where AI can really shine from a base model or data aggregator.”
for central banks,” Haghighi said. “It’s something The warning reflected Gensler’s work as global
rather tedious, and it’s something that you can economics and management professor at the
train AI to do well and do better and faster than Massachusetts Institute of Technology, where he
humans.” published a 2020 paper with Lily Bailey on deep
The European Central Bank (ECB) is using AI learning. That subset of AI offers “previously
for applications such as automating classification unseen predictive powers enabling significant
of data from 10 million business and government opportunities for efficiency, financial inclusion,
entities, scraping websites to track product prices and risk mitigation,” they wrote. But they cau-
in real time. It is also using the technology to help tioned that financial regulations rooted in ear-
bank supervisors find and parse news stories, super- lier eras “are likely to fall short in addressing the
visory reports, and corporate filings. systemic risks posed by broad adoption of deep
With the data universe growing exponentially, learning in finance.”
cleaning it up to be intelligible is a key issue, espe-
cially for unstructured data, said Myriam Mou- ‘Polycrisis’ factor
fakkir, the ECB’s chief services officer. AI can Another danger is that AI tools may exacerbate
help humans make important distinctions. The a crisis, whatever the cause, because they are
ECB is also exploring large language AI models trained on past data that may not reflect reality in
to help write code, test software, and even help an unprecedented situation, according to Anselm
make public communications easier for people Küsters, head of the digitalization and new tech-
to understand. nologies department at the Centre for European
Policy in Berlin. Küsters has cited the term poly-
Financial stability crisis, popularized by fellow economic historian
London School of Economics researcher Jon Dan- Adam Tooze, referring to interaction of different
ielsson, who studies how AI affects the financial shocks that together add up to be worse than the
system, sees the technology’s capabilities on a sum of their parts.
continuum from basic to advanced, he said. On Increased use of opaque AI applications “creates
the basic side, there’s chess, with pieces on a board new systemic risks,” as they can quickly amplify
and rules known to all. AI easily beats humans negative feedback loops, Küsters wrote, urging the
there, but its advantage diminishes with complex- European Parliament to “focus on the additional
ity. People in unexpected situations can draw on a risks of algorithmic prediction arising in crises.”
range of knowledge to make better-informed deci- Such questions posed by the rapidly evolv-
sions, from economics and history to ethics and ing technology will confront central bankers and
philosophy. And this, he said, is where humans other policymakers in coming years as benefits and
beat AI—for now. threats become clearer.
AI is already making important financial deci- “We are not yet at the point where we know what
sions, such as handling credit card applications, and makes sense for central bankers,” the ECB’s Mou-
it’s making rapid inroads in the public and private fakkir said. “We are at the beginning.” F&D
sectors. The technology can help ensure that banks
don’t misbehave by, for example, taking advantage jeff kearns is on the staff of Finance &
of clients or allowing fraud or money laundering, Development.
he said. At the same time, such expanded uses may
introduce danger, he said.
DECEMBER 2023 41
F&D Artificial Intelligence
TECHNOLOGY FOR
DEVELOPMENT
Daniel Björkegren and Joshua Blumenstock
A
s artificial intelligence (AI) reshapes develop-
ing economies, it raises familiar risks of disrup-
tion, misinformation, and surveillance—but
also promises many potential benefits. Recent
examples illustrate how AI-based technologies
can target aid and credit better and improve access to tai-
lored teaching and medical advice. But balancing these
risks and opportunities means more than just plug and
play of existing technology—it calls for local innovation
and adaptation.
Most recent advances in artificial intelligence origi-
nated in wealthy nations—developed in those countries
for local users, using local data. Over the past several
42 D E C E M B E R 2023
G ER AL D AN DER SO N/ AN ADOLU AGE N CY V IA G ET T Y IM AG ES
Artificial Intelligence F&D
Kenyan youth years, we have conducted research with partners in veloped in aerial imagery and mobile subscribers
attend a low-income nations, working on AI applications for with low balances on their phones. Targeting based
robotics and
coding class those countries, users, and data. In such settings, on these signatures helped ensure that cash trans-
to learn future AI-based solutions will work only if they fit the local fers reached people with the greatest need (Aiken
technologies. social and institutional context. and others 2022).
In Togo, where the government used machine This application worked in Togo only because
learning technology to target cash aid during the the government, in collaboration with research-
COVID-19 pandemic, we found that adapting AI to ers and nonprofit organizations, customized the
local conditions was the key to successful outcomes. technology to meet local needs. They built a sys-
The government repurposed technology originally tem for distributing mobile money payments that
designed to target online advertising to the task of worked for all mobile subscribers, adapted existing
identifying the country’s poorest residents. Using machine learning software to target cash transfers,
AI, the system processed data from satellites and and interviewed tens of thousands of beneficiaries
mobile phone companies to identify signatures of to ensure that the system reflected the local defini-
poverty—such as villages that appeared underde- tion of poverty. And even then, the AI-based solu-
DECEMBER 2023 43
F&D Artificial Intelligence
investment in knowledge
The AI-based program also raised another con-
cern: algorithms that perform well in a laboratory
infrastructure, especially
may not be reliable when deployed for consequen-
tial decisions on the ground. For instance, in an
in developing economies,
aid-targeting system like the one in Togo, people
might adapt their behavior to qualify for benefits,
underrepresented.”
phone behavior (Björkegren and Grissen 2020).
For example, in Kenya over a quarter of adults have
taken out loans using their mobile phones. But if
those with more Facebook friends are likelier to be
approved for a loan, some applicants may consider
adding friends quickly. Ultimately, this can make
it hard for systems to target the intended people.
In a study with the Busara Center in Kenya, we
found that people were able to learn and adjust their
smartphone behavior in response to such algorith-
mic rules (Björkegren, Blumenstock, and Knight,
forthcoming). We showed how a proof-of-concept
adjustment to the algorithm, which anticipates
these responses, performed better. However, tech-
nology alone cannot overcome problems that arise
during implementation; much of the challenge of
building such systems is ensuring that they are reli-
able in real-world conditions.
On the other hand, some systems require adap- algorithms with negative numbers and fractions to
tation before they will be useful. For instance, in low-income residents. But our team found simpler
many lower-income countries, teachers must han- ways to communicate these concepts. It was clear
dle large classes with limited resources. In Sierra when people responded to the algorithm that they
Leone, a local partner piloted an AI chatbot system grasped the concept. Still, complex AI systems are
for teachers, called TheTeacher.AI, which is simi- difficult to understand, even for AI researchers.
lar to ChatGPT but tailored to local curriculum and Some applications don’t require that users know
instruction and accessible even when internet con- how algorithms work. For instance, Netflix movie
nections are poor. In the pilot phase, many teachers recommendations can benefit users even if they do
couldn’t phrase questions in a way that yielded use- not understand how the algorithm selects content
ful answers, but a small group began to use the sys- it thinks they will like. Likewise, in a humanitar-
tem regularly to help with teaching concepts, plan- ian crisis, policymakers may deem it acceptable to
ning lessons, and creating classroom materials (Choi use an inscrutable “black box” algorithm, as Togo’s
and others 2023). It took training and experimenta- government did in response to the COVID-19 crisis.
tion for teachers to use it in practice. Uses of AI may Transparency is sometimes critical. When tar-
not be immediately obvious to those who stand to geting social protections in nonemergency settings,
benefit; discovering the many uses will depend on explaining eligibility criteria to potential benefi-
trial and error and sharing applications that help. ciaries is essential. This is easier said than done:
scores of interviews and focus groups showed us
Communication barriers how norms and values around data and privacy are
Grasping the potential of AI is likely to be harder for fundamentally different in a setting such as rural
people in lower-income countries, where literacy and Togo than in wealthy nations, where AI-based sys-
numeracy are lower and residents are less familiar tems are more common. For instance, few people
with digital data and the algorithms that process this we spoke to were worried about the government or
information. For instance, in our field experiment in companies accessing their data (a dominant con-
Nairobi, Kenya, we found it difficult to explain simple cern in Europe and the United States), but many
44 D E C E M B E R 2023
Artificial Intelligence F&D
and then validate the response than to collate infor-
mation from several online resources.
Some AI systems will, however, require invest-
ment in knowledge infrastructure, especially in
developing economies, where data gaps persist and
the poor are digitally underrepresented. AI models
there have incomplete information about the needs
and desires of lower-income residents, the state of
their health, the appearance of the people and vil-
lages, and the structure of lesser-used languages.
Gathering these data may require integrating
clinics, schools, and businesses into digital record-
keeping systems; creating incentives for their use;
and establishing legal rights over the resulting data.
Further, AI systems should be tailored to local
values and conditions. For example, Western AI
systems may suggest that teachers use expen-
sive resources such as digital whiteboards or dig-
ital slide presentations. These systems must be
adjusted to be relevant for teachers lacking these
resources. Investing in the capacity and training of
local AI developers and designers can help ensure
that the next generation of technical innovation
better reflects local values and priorities.
Artificial intelligence promises many useful
applications for the poor across developing econ-
omies. The challenge is not in dreaming big—it’s
easy to imagine how these systems can benefit the
poor—but in ensuring that these systems meet peo-
ple’s needs, work in local conditions, and do not
Juliana Rotich wondered if and how such information would be cause harm. F&D
works on a shared with their neighbors.
laptop at the
i-Hub technology As AI is more commonly deployed, populations daniel björkegren is an assistant professor
innovation must understand its broader societal effects. For at the Columbia University School of International
center in instance, AI can generate provocative photographs and Public Affairs. Joshua blumenstock is
Nairobi, Kenya.
that are entirely false and robocalls that mimic a chancellor’s associate professor at the University of
voices. These rapid changes will affect how much California, Berkeley, School of Information and the
people should trust information they see online. Goldman School of Public Policy.
Even remote populations must be informed about
these possibilities so that they are not misled—and references
to ensure that their concerns are represented in the Aiken, Emily, Suzanne Bellue, Dean Karlan, Chris Udry,
and Joshua E. Blumenstock. 2022. “Machine Learning
development of regulations.
and Phone Data Can Improve Targeting of Humanitarian
WAL DO SW IE G E RS /B LOOM B ERG V IA GE T T Y IMAGE S
DECEMBER 2023 45
F&D Artificial Intelligence
AI IN PRACTICE
O
n the ground and in the air, big changes are brew-
ing, powered by the latest technological dis-
ruption, artificial intelligence. AI’s potential to
transform society is vast; here we examine how
it is impacting three key sectors—agriculture,
medicine, and defense.
AI is revolutionizing these (and many other) fields, from
optimizing crop yields and improving precision healthcare
to enhancing military capabilities and national security.
The following real-life examples from around the world
highlight some of the technology’s tangible benefits but
also raise questions about the ethical considerations, pol-
icy lags, and training gaps associated with the integration
of AI into critical industries. They attempt to shed light on
the boundless possibilities and challenges that lie at the
intersection of technology and human progress.
46 D E C E M B E R 2023
Artificial Intelligence F&D
Growing More with Less ity sectors are food, agriculture, and digital indus-
In the battle against hunger, AI can help fewer tries. Those are woven together in government-run
farmers generate more food Smart Farmer and Young Smart Farmer programs
that encourage growers to adopt precision agricul-
Robert Horn ture by connecting with new technologies. These
include AI-controlled drones and software for
artificial intelligence (ai) is making Tea plantation intelligent and targeted spraying to increase yields
its mark on food and agricultural production in Hangzhou, while protecting the environment and ecosystem.
China: AI-
chains. The groundbreaking technology is already driven smart Food and agriculture have long been sources of
in use to engineer new varieties of climate-resis- agriculture strength for Thailand. The Southeast Asian king-
C OSTF OTO /NU R PHOTO V IA GE T T Y IMAG ES
tant rice; provide data on soil; guide drones that provides dom of 70 million people is the world’s 15th largest
tremendous
precision-spray fertilizers and pesticides; and sort, potential for food exporter and the only net food exporter in Asia.
inspect, and grade produce. “AI-driven smart agri- boosting food With a projected $44.3 billion in shipments this year,
culture provides tremendous potential for boosting security and to Thailand plays a crucial role in regional and global
reduce or even
food security and to reduce or even end hunger in end hunger in food security and the campaign to end hunger.
many regions of the world,” said Channing Arndt, many regions. But that campaign, despite impressive gains
of the Consultative Group on International Agri- during the past decade, has recently suffered set-
cultural Research, a global research partnership. backs. The pandemic, war in Ukraine, and result-
Policymakers in Thailand agree. In 2014, they ing disruptions left an estimated 735 million people
unveiled Thailand 4.0, a 20-year national strat- (9.2 percent of the global population) undernour-
egy for advanced development. Among its prior- ished in 2022, according to The State of Food Security
DECEMBER 2023 47
F&D Artificial Intelligence
about risks. If the data are bad, AI’s results mation and tools that help smallholder farmers
choose the right crop varieties and precision meth-
will be bad.” ods to increase productivity and profitability. Its
portal also assists farmers with a perennial press-
ing problem: access to finance. Meanwhile, Mitr
Phol Group, the largest sugar producer in Asia, has
partnered with IBM for AI-driven data solutions for
farmers, and Chia Tai, one of Thailand’s biggest
agri-food companies, is using autonomous drones
made by XAG of China.
and Nutrition in the World, published by the United
Nations. Even food-rich Thailand saw hunger rise smarter policies
for the first time in a decade. In response, policy- But smart farming is still relatively rare. Ricult
makers in several regions are exploring how digital cofounder Aukrit Unahalekhaka said that gov-
technologies can make agriculture more productive ernment agencies trying to micromanage while
and food chains more efficient to turn the tide on working in insulated teams hinder uptake. It’s a
malnutrition and food scarcity. problem, he said, throughout the region. “Govern-
ments should be creators of policies and facilitators
using ai to fight hunger of funding for start-ups, innovators, and farmers.
Thailand is just one example of how countries are It is much more efficient to let the market work,”
using AI to combat rising hunger, food insecurity, Unahalekhaka said.
and poverty, which take a toll on economies. Under- That doesn’t always happen. Several govern-
nourished people need more public assistance and ments in Africa, another continent struggling with
are less productive workers, which can affect per hunger and food security, have passed restrictive
capita income, growth, and sometimes political drone regulations, and acquiring a license for one is
stability. At the same time, younger farmers are difficult, according to “Empowering Africa’s Food
migrating to better-paying jobs in cities, leaving Systems for the Future,” a report by the Interna-
fewer hands to produce an increasing volume of tional Food Policy Research Institute (IFPRI).
food for a growing global population. Those com- However, countries such as Kenya, Rwanda, Tan-
bined trends could spell crisis, but analysts and pol- zania, and others are devoting resources to build-
icymakers see hope in new technologies, including ing a digital ecosystem and literacy for farmers
AI, to help fewer farmers generate more food. so that they can access online extension services,
Nurturing the digital ecosystem is foundational weather forecasts, market information, and financ-
in that effort, said Krithpaka Boonfueng, executive ing. Obstacles such as connectivity and digital liter-
director of the National Innovation Agency (NIA). acy, however, remain. “While the digital revolution
In October, Thailand launched THEOS-2, the first holds immense promise for African food systems,
Earth observation satellite jointly designed by Thai addressing these challenges is pivotal for its suc-
and British engineers, which will gather data for cess,” according to IFPRI.
smart agriculture. NIA has incubator and accelera- For all the promise, some researchers warn
tor programs that help source private sector invest- about risks. If the data are bad, AI’s results will be
ment for agricultural technology start-ups to deliver bad. And AI can be programmed to increase yields
the data to the field. The Digital Economy Promo- while ignoring negative impacts on the environ-
tion Agency (DEPA), another technology arm of the ment. “AI can be fine-tuned to match your goals. It
government, manages the One Community, One is not perfect,” Unahalekhaka said, adding that he
Drone program, which has farmers in 500 commu- hasn’t seen farmers misuse AI so far. He is one of
nities sharing drone services to manage their fields. many who believe that the benefits outweigh the
“Even farmers want technology, but the technology risks and that results are likely to be positive. That’s
has to be simple enough for them to use,” said Pree- because of the motivation he believes he shares
san Rakwatin, executive vice president of DEPA, with others in agricultural technology: “We want
which helps match tech businesses with markets to make the world a better place.”
and also funds start-ups.
One of those start-ups, Ricult, is already help- robert horn is a Bangkok-based freelance
ing farmers in Pakistan, Thailand, and Vietnam. writer who previously worked at Fortune magazine,
Founded in 2015, Ricult is a dual fintech and Time magazine, and the Associated Press.
48 D E C E M B E R 2023
Artificial Intelligence F&D
productivity of a retina clinic whose patients with
diabetes were randomly assigned to the AI or the
control group.
The result? When the AI tool was used, an esti-
mated 1.59 patients an hour received what was
deemed a high-quality visit, versus 1.14 in the
control group, wrote Digital Diagnostics founder
Michael Abramoff and his coauthors in an article
published in October in Nature Portfolio's npj Dig-
ital Medicine journal.
The test showed that LumineticsCore could
help more people get screened for eye damage
from diabetes, even in developing economies, said
Abramoff, who is also a professor of ophthalmology
and engineering at the University of Iowa.
He distinguishes between what he calls “impact
AI” in medicine and “glamor AI,” meaning prod-
ucts that garner tantalizing headlines but do not yet
show hard evidence of benefit to patients.
“We like what we now call impact AI,” which is
shown to help improve people’s health, Abramoff
said. “I’m an engineer, so I love technology, but
we shouldn’t be paying too much for it if it doesn’t
improve outcomes.”
And there’s a need for vigilance with AI because
its application in medicine has already shown
potential to harm as well as help people.
For example, a 2019 Science paper reported that
an algorithm widely used by large health systems
and insurers underestimated the severity of Black
patients’ illness, thus setting the stage to deny them
AI’s Healing Powers care. Researchers and policy experts have raised con-
Artificial intelligence shows promise in medicine, cerns about developing AI tools based on data drawn
but there are recognized drawbacks and risks with a tilt toward relatively wealthy people, who are
often White and have good access to health care.
Kerry Dooley Young Greater diversity is essential among the patients
whose data help train AI tools, as well as among
it’s easy to imagine the potential for A technician the people who build these products, said Jerome
artificial intelligence (AI) to help people around the scans the eye Singh, one of the advisors on a 2021 World Health
of a woman in
world live healthier lives. Kianjokoma, Kenya, Organization (WHO) guidance report on ethics and
Some already use AI to spot early signs of dis- with a smartphone governance of AI for health.
ease quickly, as recently reported in a study done in application. The “You’re going to need to have multiracial, mul-
“eyephone app”
Rangpur, Bangladesh. In this study, the nonprofit provides Kenyans ticultural coders,” Singh said. “The interpretation
Orbis International, which seeks to address pre- the chance to is quite important. AI is only as good as its coding.”
ventable causes of blindness, and local physicians get a quick That need for more diversity is one of the chief
and effective
used the LumineticsCore system, from Coralville, challenges ahead for attempts to use AI in medicine
TON Y KA RU M BA /AF P V IA GE T T Y IMAGE S
diagnosis, even
Iowa–based Digital Diagnostics. The system uses a in remote rural globally, especially in the Global South, Singh said.
special kind of camera designed to capture images areas. The need for AI may be greater in less developed
of the eyes and evaluates them with AI. economies, where the ratio of medical personnel
This product already has an impressive track to patients tends to be much higher than in afflu-
record. In 2018, it was the first AI-driven device ent areas. In the United States, there are about 36
to gain US Food and Drug Administration (FDA) doctors for every 10,000 people and in the United
clearance to check for diabetic retinopathy. In 2020, Kingdom, about 32, but in India, there are about 7
the giant US Medicare health program agreed to per 10,000 people, according to WHO data.
pay for use of the device in primary care offices. Yet these less affluent countries also face chal-
In the Bangladesh study, researchers tracked the lenges when it comes to the infrastructure and insti-
DECEMBER 2023 49
F&D Artificial Intelligence
in the past, she said. ligence at the local level—to handling more mun-
“The efficiencies that have been built now on dane tasks, such as predicting logistical challenges.
the discovery side are exponential,” Fakhouri said. A fourth use involves information warfare. This is
We’re going to see a lot coming to the market soon.” a way of leveraging AI to, in the words of Matthew
Ford of the Swedish Defence University in Stock-
kerry dooley young is a freelance journalist holm and coauthor of a book on the battlefield dig-
who specializes in health care. ital explosion, Radical War: Data, Attention and
Control in the 21st Century, “shape how narrative
construction works.”
50 D E C E M B E R 2023
Artificial Intelligence F&D
But even though the war has shown that AI can data
help armies monitor enemy movements and deliver The Ukraine war has highlighted the importance of
payloads remotely and autonomously, it has also data—the fuel that powers AI—but has also raised
accelerated shifts in battlefield dynamics. Forces soon troubling questions for policymakers and planners.
alter tactics, techniques, and procedures either to Ukraine understood early that what constituted
leverage the new technology or to mitigate its impact. data in a war had shifted. It quickly reconfigured
A failure to adapt quickly can be seized on by an a government app for filing taxes to also allow cit-
agile foe. When Russian soldiers and pilots com- izens to upload photos, videos, and other details
municated without encrypting their conversations, about Russian troops and positions to a database
Ukraine developed AI-based voice recognition and run by the military.
translation software to monitor these communica- It combined commercially available satellite
tions and extract actionable intelligence. And even images with classified data from its allies, as well
when countermeasures are adopted, each side must as from hacking into Russian surveillance cameras
be ready to rethink and enhance its technology as and from its own fleet of drones. But all this data
rapidly as the other. When Russia introduced elec- needed to be turned into actionable intelligence,
tronic jamming to thwart Ukraine’s combat drones, for which Kiev turned to private tech companies—
for example, Kyiv’s cadre of programmers devel- the most visible being Palantir, a US company spe-
oped an AI tool to help its drones evade Russian cializing in big-data analytics. Palantir’s involve-
jamming and stay locked on target. ment extended the role a private company might
play in processing sensitive data, especially during
unpiloted drones a war. Its chief executive, Alex Karp, is on record as
This technological arms race is strikingly different saying the company is responsible for most of the
from how many military thinkers saw the deploy- targeting in Ukraine. According to CSET’s Kahn,
ment of AI. For one thing, the principles behind “It’s almost like a full service they provide, which I
unpiloted aerial vehicles, or UAVs, have not changed think has proved invaluable.”
significantly since the 1990s. But in Ukraine the What hasn’t been fully considered, at least pub-
range of drones, and their capabilities, has evolved licly, are the implications. Private companies, says
rapidly, largely by coupling them with continuous the Swedish Defence University’s Ford, are going to
advances in AI. While the military-grade Turk- be crucial, because they are the only organizations
ish-made Bayraktar TB2 drone played a key role in that can develop the kind of AI armed forces can
Ukraine’s defense in the early months of the 2022 use. But, he asks, “Once it’s out there, where does
invasion, it became less useful as Russia upgraded it go next? How’s it going to be controlled, shaped,
its air defense and electronic warfare capabilities. or directed?”
With more permanent battle lines drawn later in
the year, Ukraine pushed its drone makers to adapt. digital battlefield
The result has been a succession of improved and The war also introduces another aspect of AI and
diverse devices. In September, for example, Kyiv data. “The Ukraine-Russia war is the most doc-
approved the deployment of homegrown Saker umented war in history,” says Andrew Hoskins,
Scout drones, which can detect enemy targets professor of global security at the University of
often missed by the human eye, even when hid- Glasgow and Ford’s Radical War coauthor. Tele-
den under camouflage. gram, the social media platform now used by
This emphasis on rapid evolution has helped three-quarters of Ukrainians and well over a third
change thinking among military strategists, says of Russians to share videos and photos as the war
Lauren Kahn, senior research analyst at George- plays out in front of them, “is the digital battle-
town University’s Center for Security and Emerging ground of this war,” he says.
Technology (CSET). Despite excitement about AI That information is not being uploaded only to
in military circles since 2021, if not earlier, practical army and intelligence servers, but also to NGOs
examples were either hypothetical or project based. and investigators mining it to catalog human
“That changed after Ukraine,” she says. rights abuses for future war crimes trials. AI, too,
Planners began to see that AI was not just a box is improving what can be seen and extracted, says
to tick but raised a series of searching questions Hoskins. When you apply AI to these archives,
about what would make it useful: data, knowledge “you start to find things you never anticipated.” F&D
about your own side and the other, testing and eval-
uation procedures. The creative way Ukraine has jeremy wagstaff is a technology and media
developed drone technology is something “no one consultant and former journalist at the BBC,
could have imagined,” she said. Reuters, and the Wall Street Journal.
DECEMBER 2023 51
F&D AI Glossary
AI LEXICON
Artificial
Intelligence
A field of computer science
that focuses on building
systems to imitate human
behavior and demonstrate
machine intelligence.
Deep Learning
A subset of machine
learning that uses large
Bias multilayered (artificial)
deep neural networks
A phenomenon that occurs when an AI system produces that compute with
results that are systematically unfair or inaccurate continuous (real-number)
due to erroneous assumptions or influences in the machine representations, a little
learning process. Bias in AI can have negative impacts like the hierarchically
on individuals and society, such as discrimination, organized neurons in
misinformation, or loss of trust. There are different the human brain. It is
types and sources of bias in AI, such as data bias, especially effective at
algorithm bias, human bias, and societal bias. learning from unstructured
data such as images, text,
and audio.
52 D E C E M B E R 2023
AI Glossary F&D
A phenomenon in which
an AI system produces
outputs that are not based Neural Unsupervised
on reality or the given
context. For example, an AI Network Learning
chatbot might make up facts
or stories, or an AI image A computational model A type of machine learning
recognition system might inspired by the structure in which algorithms learn
see objects or patterns and function of biological patterns from unlabeled
that are not there. neurons. data, without any human
guidance or feedback.
SOURCES: Coursera. 2023. “Artificial Intelligence (AI) Terms: A to Z Glossary.” Mountain View, CA; Manning, Christopher. 2020. “Artificial Intelligence Definitions.” Stanford
University Human-Centered Artificial Intelligence glossary, Stanford, CA; and McKinsey Digital. 2023. “The Economic Potential of Generative AI: The Next Productivity Frontier.”
DECEMBER 2023 53
F&D Feature
China’s Bumpy
Path
Eswar Prasad
C
hina’s economic performance has been work, a market-oriented economy, and a dem-
stellar over the past three decades, with ocratic and open system of government. Until
remarkable and persistent high growth the COVID-19 pandemic rocked it back on its
that lifted the economy from low-income heels, the Chinese economy powered through
to upper-middle-income status. Measured periods of domestic and global turmoil seem-
at market exchange rates, China’s GDP was $18.3 ingly unscathed.
trillion in 2022, 73 percent of the GDP of the United But detractors have long argued that China’s
States and 10 times more than the 7 percent of economic collapse was imminent, pointing to
US GDP it registered in 1990. China’s per capita numerous fragilities. The country’s growth has
income is now roughly $13,000, approximately 17 been powered by investment in physical capital,
percent of US per capita income—compared with especially real estate, that has been financed by
less than 2 percent in 1990. Over the past decade an inefficient banking system. With domestic
and a half, China has been the main driver of the debt levels high and rising, the property mar-
world’s economic growth, accounting for 35 percent ket unraveling, and the labor force shrinking,
of global nominal GDP growth, while the United some analysts say the day of reckoning has
States accounted for 27 percent. finally arrived.
China accomplished this without many attri- They are likely wrong. Unbalanced reforms
butes that economists have identified as being that have kept the institutional structure weak,
crucial for growth—such as a well-functioning a schizophrenic approach to the role of the
financial system, a strong institutional frame- market versus that of the state, and strains in
DECEMBER 2023 55
F&D Feature
56 D E C E M B E R 2023
Feature F&D
nology and hi-tech products, as well as to mar- major inefficiencies and waste because of a broken
kets for its exports. Moreover, the government’s system of allocating capital.
recent crackdown on private firms in sectors such How debt and assets are distributed through-
as technology, education, and health has had a out the economy matters. Tumbling house prices
chilling effect on entrepreneurship. have caused several major property developers,
such as Country Garden and Evergrande Group,
Potential pitfalls to run into financial trouble recently, and many
There are concerns that China’s economy is others are similarly exposed—with high debt and
headed for a crash similar to those experienced vulnerable balance sheets. So are some of the
by other high-flying Asian economies—such as financial institutions that lent to them. But a sys-
Malaysia and Thailand. China’s overall debt has temic meltdown is not in the cards. Most major
been a significant concern for many years. Over Chinese banks are under state control and can
time debt has risen relative to the size of the econ- provide infusions of cash to troubled corpora-
omy—although gross debt levels are not out of tions, even if that only pushes problems off into
line with those of other major economies, such the future. Stumbles are inevitable as China tries
as the United States and Japan. Moreover, pub- to give market forces freer rein, but the govern-
lic borrowing as a percentage of nominal GDP is ment has enough control and resources to prevent
lower in China than in other major economies. broader financial crashes.
China has a high level of corporate debt—about
131 percent of GDP. But most of it is denominated External risks
in China’s own currency and owned by domes- Many emerging market economies have run into
tic banks and investors, which presents less of a distress from high levels of external debt, partic-
threat than were the debt owed to foreign inves- ularly foreign currency debt, which can cause bal-
tors and denominated in foreign currencies, such ance sheet problems when a country’s economy
as the US dollar. and exchange rate deteriorate simultaneously. But
There are, however, specific sectors in which China’s external debt is estimated to be a modest 16
the concentration of debt could be a problem— percent of GDP, and less than half of it is denomi-
especially the real estate sector. Real estate nated in foreign currencies.
investment has become a bulwark of the econ- Still, economic and political uncertainty have
omy, helping to keep growth on an even keel when created concerns about capital flight, which could
other sectors floundered. Local government offi- bring down the financial system and cause the cur-
cials are eager to sell land to developers, boosting rency’s value to crash. But this is an unlikely sce-
public revenues and enabling a range of govern- nario, because much of the banking system is state
ment expenditures. So a fall in real estate prices— owned and the government would probably back all
or the emergence of other factors that restrain deposits in the event of financial panic. Moreover,
real estate activity—could have knock-on effects because the government directly controls much of
across other sectors, local government finances, the banking system, it can choke off the conduits
and even household wealth. for large capital outflows.
Household exposure to the real estate sector has Although there have been reforms in recent
created additional vulnerabilities that could affect years, many of them were related to the financial
economic and social stability. Easier access to res- sector and capital markets, with far fewer in other
idential mortgages, which the government encour- areas, such as state enterprises and the institutional
aged, boosted housing demand and contributed to framework. This lack of balance creates risks.
a surge in household debt, from about 30 percent of The government seems to have grasped the
GDP a decade ago to more than 60 percent. Prop- need for financial sector reforms and liberaliza-
erty has also become a mainstay of Chinese house- tion to promote better resource allocation. Fixing
hold wealth. Households are exposed in multiple the financial system is not just about managing
ways to house price fluctuations. Still, total house- risks and avoiding disaster but also about allocat-
hold debt is less than total household deposits in ing capital to the more productive, dynamic, and
the banking system. employment-generating parts of the economy.
Because debt accumulation in China has been China’s financial system is still dominated by
financed mostly by domestic savings, overall finan- banks, whose loan portfolios are concentrated in
cial risk is limited. The state owns many of the key the state enterprise sector. Fixing the banking sys-
creditors and debtors, which means a financial tem requires recognizing and removing bad loans
shock is unlikely to set off a financial crisis or a from banks’ balance sheets, as well as reform of the
collapse in growth. The more pertinent issues are state enterprises themselves, including weaning
DECEMBER 2023 57
F&D Feature
them off dependence on bank credit. enough resources and policy space to cope with
In recent years, as it dealt with episodes of hous- some of those transitional risks, but its actions
ing market and stock market volatility, the govern- and attempts to intervene directly in markets at
ment often found itself caught in a schizophrenic difficult times might exacerbate problems, with
effort to balance maintaining confidence in the long-lasting consequences.
market with allowing the market to discipline
itself—which had the perverse effect of heightening What the future holds
market turbulence. This on-off approach to inter- The Chinese government has shown an uncanny
vention has sometimes injected a strong dose of ability to manage the severe economic and finan-
uncertainty on top of already fragile investor sen- cial stresses that have built up from the highly inef-
timent and added to market volatility. ficient and risky growth model it had embraced. At
Moreover, market-oriented reforms can back- various points, the government has maneuvered
fire, adding to volatility and generating more risks the economy around the seemingly inevitable
if they are not accompanied by broader reforms. prospects of a banking crisis, massive currency
China needs more transparency in its policy- devaluation, housing market meltdown, and eco-
making process, better corporate governance nomic collapse.
and accounting standards, and more operational Yet each of these near misses has exacted a toll:
independence for the central bank and regulatory a huge buildup in domestic debt, loss of $1 tril-
authorities to supplement its financial and other lion in foreign exchange reserves during 2015–16,
market-oriented reforms. and highly volatile prices of stocks, property, and
The government has rightly encouraged the other assets.
development of stock and corporate bond mar- The government now faces a number of policy
kets. But it has done little to improve corpo- dilemmas: how to continue reducing debt while
rate governance of Chinese companies or their maintaining growth, how to reduce energy-inten-
accounting and auditing standards. The resulting sive production while the economy continues to
opacity has contributed to large fluctuations in rely on heavy industry, how to get markets to exert
stock and bond markets, because investors have financial discipline even as the government tries
limited information about the companies they are to strengthen state control, how to restrain wealth
investing in, leading them to follow and exacer- inequality while relying on the private sector to gen-
bate market swings. erate more wealth, how to encourage private sector
Reconciling the government’s two contradic- innovation while cutting successful private enter-
tory impulses—more freedom for markets but prises down to size.
with a heavy hand of government intervention The government’s attempts to resolve these
to maintain “stability and order”—poses difficult inherently contradictory impulses in the guise of
challenges. Implementing even well-intentioned market-oriented socialism will inevitably lead to
reforms in an economy with rampant inefficien- further stumbles and accidents. Its policy approach,
cies involves transitional risks that might manifest although driven by the right objectives, could gen-
in financial and economic volatility, especially if erate more uncertainty and volatility in the short
the government does not clearly communicate its run, which in turn could reduce public support for
policy intentions and leaves households and busi- much needed reforms to bolster long-term produc-
nesses guessing. So far, the government has had tivity and growth.
The underpinnings of China’s growth seem frag-
ile from historical and analytical perspectives. Even
if no crises materialize, unfavorable demographics,
high debt levels, and an inefficient financial system
will constrain China’s growth. Yet, if the govern-
ment plays its cards right, one could equally well
envision a more benign future for the Chinese econ-
“Reconciling the government’s two contradic omy—with moderate growth that is more sustain-
tory impulses—more freedom for markets but able from an economic, social, and environmental
perspective. F&D
with a heavy hand of government intervention eswar prasad is a professor in the Dyson
to maintain ‘stability and order’—poses School at Cornell University, senior fellow at the
Brookings Institution, and author of The Future
difficult challenges.” of Money.
58 D E C E M B E R 2023
Feature F&D
History’s Inflation
Lessons
People lining
up for gas
coupons in the
Hanover Street
Post Office
in Liverpool,
United Kingdom,
November 29, 1973
I
n the early 1970s, conflict in the Middle East set
A STUDY OF 100 INFLATION SHOCKS off a spike in oil prices that left central banks
around the world scrambling to control inflation.
SINCE THE 1970S PROVIDES VALUABLE After a year or so, oil prices stabilized and infla-
POINTERS FOR POLICYMAKERS TODAY tion started to retreat. Many countries believed
they had restored price stability and loosened pol-
icy to revive their recession-hit economies only to
Anil Ari and Lev Ratnovski see inflation return. Could history repeat?
World inflation reached historic highs in 2022
after Russia’s invasion of Ukraine triggered a terms-
W ILLI AM S/M IR RO RP IX/G E T T Y I MAG E S
DECEMBER 2023 59
F&D Feature
60 D E C E M B E R 2023
Feature F&D
The ultimate prize
Fighting inflation is difficult. But it is important to
recognize the benefits of price stability. Historically,
countries that resolved inflation had lower eco-
nomic growth in the short term than those that did
not. But this relationship reversed over the medium
and long term. Five years after the inflation shock,
countries that resolved inflation had higher growth
and lower unemployment than economies that
allowed inflation to linger.
The economics behind this finding are intui-
tive. There is a trade-off between bringing inflation
down on one hand and achieving higher growth and
lower unemployment on the other. But this trade-
off is temporary: growth recovers and jobs are cre-
ated once inflation is brought under control.
By contrast, leaving inflation unresolved comes
with its own costs of macroeconomic instability and
inefficiency. These costs accumulate for as long as
inflation remains high. Consequently, cumulative
welfare losses from unresolved or permanently
high inflation dominate over the medium to long
term (Chart 3). Countries that allow inflation to lin-
ger ultimately pay a higher price.
Central bankers are on the front line of the fight
against inflation and should pay the most atten-
tion to these lessons. But governments must not
make the task of monetary authorities harder by
adding to price pressures with loose fis-
With the right policies in place, countries could cal policy. To make fiscal support during
resolve inflationary pressures sooner than in the past. “Countries that a cost-of-living crisis less inflationary,
But it won’t be easy. Conditions in the labor mar-
ket in particular require close attention. In many allow inflation to governments should target relief to the
most vulnerable, where it will alleviate
countries, workers’ wages have fallen in real infla-
tion-adjusted terms and may need to rise again to
linger ultimately suffering most.
The past is never a perfect guide to the
catch up with higher prices. Yet wage growth could
fuel inflation if it is too high and could lead to per-
pay a higher price.” present, because no two crises are pre-
cisely alike. All the same, history offers
nicious wage-price spirals. clear lessons to policymakers today.
Historically, countries that resolved inflation Fighting inflation is a marathon, not a sprint. Pol-
successfully tended to have lower nominal wage icymakers must persevere, demonstrate policy
growth. Importantly, this did not translate into credibility and consistency, and keep their eyes
lower real wages and a loss of purchasing power, on the prize: macroeconomic stability and stron-
because lower nominal wage growth was accompa- ger growth brought about by returning inflation
nied by lower price growth. The implication for pol- firmly to target. If history is a guide, inflation’s
icymakers here is to remain focused on real wages, recent decline could be transitory. Policymakers
not nominal wages, when responding to develop- would be wise not to celebrate too soon. F&D
ments in the labor market.
Countries that resolved inflation successfully anil ari and lev ratnovski are economists
were also better at maintaining external stability. in the IMF’s Strategy, Policy, and Review
Free-floating currencies were less likely to depre- Department and European Department, respectively.
ciate sharply, and currency pegs were more likely
to survive. This is not a call for currency interven- This article draws on IMF Working Paper 2023/190,
tion. Instead, it appears that countries’ success in “One Hundred Inflation Shocks: Seven Stylized
fighting inflation—through tighter monetary policy Facts,” by Anil Ari, Carlos Mulas-Granados, Victor
and greater policy credibility—was instrumental in Mylonas, Lev Ratnovski, and Wei Zhao.
shoring up exchange rates.
DECEMBER 2023 61
F&D Feature
A Critical
Matter
Christopher Evans, Marika Santoro, and Martin Stuermer
A
scramble by competing powers to secure stra-
tegic minerals could add to price pressures and
increase the costs of the climate transition. New
trade restrictions in commodity markets more
broadly have doubled since Russia’s invasion of
Ukraine as producers impose curbs on shipments. Criti-
cal minerals used to make everything from electric vehi-
cles (EVs) to solar panels and wind turbines are highly
vulnerable to more severe trade restrictions. A slide
toward opposing trading blocs could substantially delay
the energy transition.
Even without the added complication of geopolit-
DAV ID MC NE W /G E T T Y IMAG ES
62 D E C E M B E R 2023
F IRST NA ME L AST
Slug F&D
DECEMBER 2023 63
F&D Feature
up prices and could make these minerals as Now policymakers debate the future of global-
important as crude oil for the world economy over ization. They worry about the fragmentation of the
the next two decades (Boer, Pescatori, and Stuer- world economy and the flouting of global trade
mer, forthcoming). rules. Trade interventions are on the rise, in the
Why are critical mineral markets particularly form of industrial policies and subsidies, import
vulnerable in the event of fragmentation? And what restrictions based on national security and envi-
could be the impact on the energy transition? ronmental concerns, and export controls to punish
geopolitical rivals and ensure domestic supply. F&D
Extreme vulnerability
Minerals such as copper, nickel, cobalt, and lithium douglas irwin is the John French Professor
are critical inputs for the energy transition. They of Economics at Dartmouth College and a
are used in EVs, batteries and wiring, and renew- nonresident senior fellow at the Peterson Institute for
able-energy technologies such as solar panels and International Economics.
wind turbines. A typical EV battery pack, for exam-
ple, needs about 8 kilograms of lithium, 35 kilograms references
of nickel, and 14 kilograms of cobalt. Charging sta- Antràs, Pol, Teresa C. Fort, and Felix Tintelnot. 2017. “The
tions require substantial amounts of copper. Margins of Global Sourcing: Theory and Evidence from
U.S. Firms.” American Economic Review 107 (9): 2514–64.
Critical minerals are extremely vulnerable in the
Grossman, Gene M., Elhanan Helpman, and Hugo Lhuillier.
event of trade disruptions because their global pro-
2023. “Supply Chain Resilience: Should Policy Promote
duction is highly concentrated. Two-thirds of the International Diversification or Reshoring?” Forthcom-
world’s cobalt is mined in the Democratic Repub- ing in the Journal of Political Economy.
lic of the Congo alone. The top three producers of
nickel and lithium control more than 60 percent
of supply. Crude oil production is, by comparison,
much more diversified (Chart 2).
The combination of concentrated supply and
widespread demand has led to extensive commod-
ity trading. Many countries rely heavily on imports
from only a handful of suppliers. To make matters
worse, mining production can be difficult to relo-
cate. Even where there are deposits, it takes time
and expensive investment to extract them from the
ground. Minerals are often hard to substitute. For
example, lithium is essential for many EV batteries.
As a result, demand for them responds only slowly
when prices rise amid shortages.
This trifecta of high concentration of production
and low reactivity of supply and demand makes
critical minerals for the energy transition highly
vulnerable in the event of trade restrictions.
Transition delay
How would more severe fragmentation of critical
mineral markets affect the energy transition? For
illustrative purposes, a team of IMF researchers
divided the markets for four critical minerals into
two hypothetical blocs that refuse to trade with
each other, along the lines of a 2022 UN vote on
Ukraine.
Results show that the inability of the hypothet-
ical China-Russia+ bloc to import copper, nickel,
lithium, and cobalt from mining countries such
as Chile, the Democratic Republic of the Congo,
and Indonesia would lead to an additional price
increase of 300 percent, on average. Acquiring
minerals would be more expensive, which would
64 D E C E M B E R 2023
Feature F&D
on enhanced World Trade Organization rules on
export restrictions and tariffs as well as discrimi-
natory subsidies would be the best solution.
If full cooperation is impossible, multilateral
efforts should prioritize establishing a “green cor-
ridor,” consisting at a minimum of agreement to
maintain the free flow of critical minerals and not to
discriminate between firms from different countries.
An international initiative to improve data shar-
ing and standardization in mineral markets could
also reduce market uncertainty. The international
community should establish an institution or plat-
form, similar to the International Energy Agency
or the Food and Agriculture Organization, focused
solely on critical minerals.
Individual countries can take proactive steps, too.
Strategies could include diversifying sources of com-
modity supplies; greater investment in mining, explo-
ration, and storage; and critical mineral recycling.
Industrial policies, meanwhile, must be
designed carefully to ensure equal treatment of
firms across competitive markets to prevent adverse
cross-country spillovers, minimize distortions and
inefficiencies, and mitigate fiscal risks and harmful
political economy outcomes. “Friend-shoring” pol-
icies and local-content provisions can also distort
markets and raise costs. Developing a framework
for international consultation on friend-shoring
could help identify negative cross-border spillovers
lead to lower investment in solar panels and wind and mitigate adverse consequences.
turbines and fewer EVs. Fragmentation in critical mineral markets could
In the hypothetical US-Europe+ bloc, mean- make the clean energy transition more costly and
while, fragmentation would cause an oversupply of potentially delay much-needed policies to mitigate
most of these mined minerals. However, the bloc’s climate change. Multilateral cooperation on trade
use of minerals would be constrained by the length policies and more data sharing would thwart addi-
of time it takes to scale up refining capacity. Frag- tional obstacles to a cleaner global energy system.
mentation, therefore, generates only small gains in Critical minerals may someday be as important to
the US-Europe+ bloc by 2030: the bloc would pro- the world economy as oil is today. We need a better
duce slightly more EVs, but there would be no gains understanding of their complex value chains. F&D
in renewable-energy capacity.
Decarbonizing the global economy would be christopher evans and martin
more difficult if the market for minerals were stuermer are economists in the IMF’s Western
fragmented. On balance, global net investment Hemisphere Department and Research Department,
in renewable technology and production of EVs respectively. marika santoro is a senior
would be about 30 percent lower, if greenhouse economist in the Strategy, Policy, and Review
gas emissions are used as weights to aggregate Department.
region-specific results (Chart 3). This measure
accounts for the greater emissions intensity of This article draws on Chapter 3 (“Fragmentation and
activity in the China-Russia+ bloc and hence the Commodity Markets: Vulnerabilities and Risks”) of the
greater effort needed to achieve global emissions IMF’s October 2023 World Economic Outlook.
mitigation goals.
reference
International initiatives Boer, Lukas, Andrea Pescatori, and Martin Stuermer.
Forthcoming. “Energy Transition Metals: Bottleneck
Multilateral cooperation is essential to prevent
for Net-Zero Emissions?” Journal of the European
vicious spirals where countries impose trade restric- Economic Association.
tions as a risk management tool. An agreement
DECEMBER 2023 65
F&D People in Economics
PO RT E R G IF F OR D
66 D E C E M B E R 2023
F&D
People in Economics
The Inequality
Economist
Bob Simison profiles Harvard’s Lawrence F. Katz,
whose research changed economists’ understanding
of economic disparity
like the rest of us, harvard labor economist Lawrence F. Katz has been “When the education
thinking about how artificial intelligence (AI) will change the future—especially system doesn’t
keep up, you have
what it will mean for inequality. Since the 1980s, he has made groundbreaking con- widening inequal
tributions to economists’ understanding of the issue and what can be done about it. ity,” Katz says.
Under one AI scenario, Katz says, the technology could help people who are
already in advanced, high-paying professions, “thereby potentially exacerbating
labor market inequality.” Under another, it might help level the playing field for
workers on the lower end of the scale.
“AI may increasingly substitute for elite expertise, making it less scarce and making
the insights from elite expert knowledge more accessible to a broader range of work-
ers,” he says. “This scenario could help middle-skill workers versus elite professionals.”
Whichever prevails, the 64-year-old Katz is likely to help lead the charge by
academic economists—many of them his own protégés—to assess AI. Certainly,
researchers will deploy his rigorous methods using big data and sophisticated anal-
ysis, reflecting his pervasive influence on economics over almost 40 years.
“He really casts a long shadow in economics,” says the Massachusetts Institute
of Technology’s (MIT’s) David Autor. “Larry was sounding the bell on the danger of
rising inequality long before anyone else did.” Autor is one of Katz’s more than 200
former doctoral students. They include two winners of the John Bates Clark Medal,
the top prize for economists younger than 40; three MacArthur Foundation grant-
ees; and dozens of tenured scholars at top universities.
Katz’s work sparked two intellectual revolutions in economics, according to a 2023
biographical sketch by Autor and Harvard’s David Deming, another former Katz stu-
dent. One was to apply economic theories of supply and demand to explain fluctua-
tions in wage inequality over time. The other was to lead large-scale field experiments
involving real people to answer big questions in social science, most prominently on
the effects over multiple generations of moving to a higher-opportunity neighborhood.
DECEMBER 2023 67
F&D People in Economics
In addition, as the editor since 1991 he began developing his data-driven a state to rebound from an unemploy-
of the august Quarterly Journal of Eco- approach to economic research when ment spike, the decline in the jobless
nomics, “Katz has shaped the agenda what’s now known as the Fisher Center rate reflects largely workers leaving the
of the economics profession over three for Real Estate and Urban Economics state rather than employers creating
decades,” Autor and Deming write. They hired him in 1979 as its first researcher. new jobs. It takes more than a decade
cite data suggesting that the QJE, as the He surveyed land-use officials from for wages to return to normal.
journal is known, has had much greater the 93 San Francisco Bay Area juris- “We found very solid patterns in
influence on economics per paper pub- dictions to collect reams of data show- the data, which gave a clear picture of
lished than any of the four other leading ing how the recently passed proper- labor mobility and regional evolutions,”
economic research publications, based ty-tax-slashing Proposition 13 was Blanchard says. The findings changed
on citations and other factors. leading to more restrictions on land the way economists think about regional
use and driving up real estate prices. policies in other places, such as Europe,
‘Driven by social problems’ The findings turned into his senior the- he says.
“Larry’s really driven by social problems,” sis and his commencement address for The other landmark 1992 paper
says his wife and frequent research col- the economics department. directly addressed income inequality
laborator, 2023 Nobel laureate Claudia In earning his PhD at MIT in 1985, between people with and without col-
Goldin, another Harvard economist. Katz dug into the mechanics of unem- lege degrees. It overturned economists’
“His passion is the underprivileged.” (He ployment. Rigorously analyzing US thinking about earnings disparity. Katz
does have another passion, she says: her and UK data, he challenged an estab- and the University of Chicago’s Kevin
champion scenting dog Pika, a 13-year- lished theory that cyclical variations in Murphy analyzed changes in US wages
old golden retriever Katz walks several joblessness grew out of shifts in labor from 1963 to 1987, tapping into a vast
times a day.) demand requiring workers to move Census Bureau dataset. They found that
His passion for the underprivileged across sectors such as manufacturing the income gap narrowed from 1970 to
came from growing up as the son of a and services. He showed that instead 1979 and widened dramatically after
Los Angeles public school psychol- it had more to do with traditional busi- 1979. The conventional wisdom at the
ogist in the 1960s. His mother, born ness cycles from aggregate demand time attributed this to rising demand for
Vera Reichenfeld in 1938 in Belgrade, shocks. He further advanced under- workers with more education. But Katz
escaped the Holocaust with her fam- standing of the job-search behavior of and Murphy showed that it also reflected
ily and grew up in Argentina and Uru- workers who were temporarily idled a sharp decline in growth of the supply of
guay. One of her teachers had studied and expected to be recalled, relying such workers relative to rising demand.
at the University of Michigan. That led on longitudinal survey data. “When the education system doesn’t
her to emigrate to Ann Arbor for college, This kind of large-scale data analy- keep up, you have widening inequal-
where she met Katz’s father. sis opened a new frontier in econom- ity,” Katz says. Those two research
As a Spanish speaker, she worked ics long before advances in computing efforts set off “a work of passion” delv-
in some of the poorest neighborhoods power enabled researchers to routinely ing into inequality over the succeeding
in LA. Katz recalls that she took cloth- crunch huge volumes of numbers. At three decades, he says. One of the most
ing and food to her schools for children that time, “public-use datasets came on important, longest-running projects
from struggling families. He and his nine-track tapes the size of deep-dish was his collaboration with Goldin on
mother also discussed the hardships pizzas, and computer time was rented their 2008 book, The Race between Edu-
of attending schools without air-con- by the processor minute,” according to cation and Technology.
ditioning and whether that put kids in Autor and Deming.
poorer schools at a disadvantage com- Katz put his approach on full display Education and inequality
pared with students in richer, air-con- in 1992, when he published two influ- The couple, who met in the late 1980s at
ditioned schools. Those encounters ential papers. In one, he collaborated the back entrance of the National Bureau
with poverty inspired Katz to focus on with the French macroeconomist Oliv- of Economic Research in Cambridge,
inequality, segregation, and race as a ier Blanchard, who later served as chief Massachusetts, started the research in
high school debater and as an under- economist of the International Mone- the early 1990s. It grew out of some of
graduate. (At the age of 85, his mother tary Fund. After joblessness nearly tri- Goldin’s preliminary work on the history
now works part-time as an actress in pled in Massachusetts between 1987 of education and its impact on wages,
Spanish- and English-speaking roles.) and 1991 as a boom in tech and finan- she says. “Larry was obsessed with
Katz likes to tell interviewers that cial services went bust, they set out to changes in the wage structure,” she says.
he chose economics because the intro- understand what happens when there’s “He was the first economist in the 1980s
ductory class during his first quarter at a regional surge in unemployment. who saw the inequality gap expanding.”
Berkeley didn’t meet until 10 a.m., and Studying 40 years of state-by-state The researchers tapped into “tons
the beginning political science course US data, Katz and Blanchard concluded of datasets” and manually tabulated
was at 8 a.m. As an undergraduate, that while it takes five to seven years for Bureau of Labor Statistics data from the
68 D E C E M B E R 2023
People in Economics F&D
early 20th century, Katz says. They dug After the LA riots over the 1991 police and another Katz protégé. Chetty has
up campaign materials created in the beating of Rodney King, “Congress felt played a leading role in studying the
1910s and 1920s by local school boards— a little responsible and passed a bill with ramifications of the Moving to Oppor-
in rural areas more than in cities—push- some money for a demonstration proj- tunity project.
ing the “high school movement” to pre- ect on neighborhoods,” Katz says. The “He is highly respected by the authors,
pare young people for better jobs. program began in 1994 in Boston, Balti- an exceptional feat,” says France’s
This gave US workers a tremendous more, Chicago, New York City, and Los Blanchard, who was coeditor of the jour-
advantage as “America educated its Angeles and included 4,604 families nal with Katz for seven years. “Strong
youth to a far greater extent than did living in public housing in some of the editors typically make many enemies.
most, if not every, European country,” country’s poorest neighborhoods. The He has not.” Katz reads and responds
they write. “By the 1930s, America was idea was to find out whether helping ran- to every paper that’s submitted, he says.
virtually alone in providing universally domly assigned families move to a bet- The QJE receives about 2,000 submis-
free and accessible secondary schools.” ter neighborhood would benefit them sions a year and publishes 48.
Widening inequality in the US by the economically. For the past 25 years, Katz has also
end of the 20th century reflected not so It didn’t, at first, Katz and other played a role as mediator in labor nego-
much the speed of technological change, researchers found. But that was only tiations and disputes between Harvard
they argue, as a shortfall in willingness part of the story. Participants did and various unions. He led what was
to continue investing in education. report improved physical and mental unofficially known as the Katz Com-
“We could have done the same thing health, and as Katz and his colleagues mittee, which in 2001 issued a report
with college and vocational educa- continued following the group, some- on outsourcing that led to a wage and
tion that we did with the high school thing unexpected emerged. Children benefits parity policy between in-house
movement,” Katz says. “We have an who were younger than 13 when their and contracted-out workers. The policy
incomplete postsecondary education families moved to safer, lower-poverty aimed to allow Harvard to use outsourc-
revolution. We left it to families to pro- neighborhoods had 30 percent higher ing for efficiency gains but not to under-
vide that.” Today he advocates invest- earnings as young adults, were more cut unionized university employees.
ing more in state universities and in likely to attend college, got into better Certainly a pillar of his legacy are the
strong vocational education and sec- colleges, and lived in lower-poverty 239 PhD economists Katz has trained.
toral employment training programs neighborhoods as adults. He maintains an up-to-date nine-page
for high school graduates. (He and col- “Little did I know I would still be list of them on his Harvard website,
laborators published a series of papers studying this more than 25 years later,” showing the year of each one’s doctor-
in the 2010s showing that employers put Katz says. ate, initial posting, and current position.
little value on degrees from costly for- The experiment has policy rami- Many of them cite him as their profes-
profit colleges.) fications today as some local govern- sional inspiration.
Fifty to 60 percent of the rise in US ments, such as Seattle’s, apply the find- “He is a prolific advisor who’s had an
wage inequality since 1980 grew out of ings to recipients of housing vouchers. enormous impact on public policy by
the slowdown in educational advances “Where you live affects how healthy you nurturing so many leading economists,”
relative to continuing growth in demand are and many other things,” Katz says. says the University of Michigan’s Betsey
for college-educated workers, which “We could do a lot more using existing Stevenson. “He was always available.
widened the pay differential between resources.” The administration of Joe He has an encyclopedic knowledge of
those with and without college degrees, Biden sought to fund a broader program, research in the field and can instantly
Katz says. Other factors include the but “it all got killed” in negotiations with tell you where your project would fit in
decline of unions, the erosion of the Congress, Katz says. the literature.”
federal minimum wage, the surge in As a graduate student, Stevenson was
executive and other top-end compen- The Katz effect doing research on happiness and eco-
sation, and the fissuring of supply chains As editor of the QJE for the past 32 years, nomics. She recalls telling Katz of her
with increased domestic outsourcing, Katz has magnified his influence on eco- finding that winning the lottery usually
greater use of the gig economy, and nomic research, other economists say. makes people happier, at least initially.
international offshoring, he suggests. Under his leadership, the journal takes on “Winning the lottery probably
In 1993, Katz became chief econo- big questions in social science and human wouldn’t make me any happier,” she
mist of the Department of Labor during welfare, extending the frontiers of eco- says he told her. “It wouldn’t help me
the administration of Bill Clinton. That nomics, according to Autor and Deming. write papers any faster.” F&D
put him in a position to help design what He pushes researchers to take risks and
other economists call one of the most follow the data where it leads, others say. bob simison is a freelance writer
important social policy experiments in “In the field, it’s known as the Katz who previously worked at the Wall
US history, the Moving to Opportunity effect,” says Harvard economist Raj Street Journal, the Detroit News,
housing mobility program. Chetty, a John Bates Clark medalist and Bloomberg News.
DECEMBER 2023 69
F&D
Straight Talk
Harnessing AI
for Global Good
B
eginning in the 18th century, the Indus- ity growth by 2–3 percentage points
trial Revolution ushered in a series of on average: some show nearly 7 per-
innovations that transformed society. centage points. Although it is difficult
We may be in the early stages of a new to gauge aggregate effect from these
technological era—the age of generative artificial types of studies, such findings raise
intelligence (AI)—that could unleash change on hopes for reversing the decline in
a similar scale. global productivity growth, which has
History, of course, is filled with examples of been slowing for more than a decade.
technologies that left their mark, from the print- A boost to productivity could raise
ing press and electricity to the internal combus- incomes, improving the lives of peo-
tion engine and the internet. Often, it took years— gita gopinath is first ple around the world.
if not decades—to comprehend the impact of these deputy managing director But it is far from certain the net
advances. What makes generative AI unique is the of the IMF. impact of the technology will be posi-
speed with which it is spreading throughout society tive. By its very nature, we can expect AI
IM F P HOTO /K IM H AU GH TO N
and the potential it has to upend economies—not to to shake up labor markets. In some sit-
mention redefine what it means to be human. This uations, it could complement the work
is why the world needs to come together on a set of humans, making them even more
of public policies to ensure AI is harnessed for the productive. In others, it could become
good of humanity. a substitute for human work, rendering
The rapidly expanding body of research on AI certain jobs obsolete. The question is
suggests its effects could be dramatic. In a recent how these two forces will balance out.
70 D E C E M B E R 2023
Straight Talk F&D
the pioneers of AI
vary both across and within countries roids. In addition, a lack of transparency
depending on the type of labor. Unlike behind this incredibly complex technol-
it poses.”
advanced economies like the US and fidential data into the public domain.
UK, with their high shares of profession- And knowing the serious concerns
als and managers, face higher exposure: about embedded bias with AI, relying
at least 60 percent of their employment on bots to determine who gets a loan
is in high-exposure occupations. could exacerbate inequality. Suffice
On the other hand, high-skill occu- it to say, without proper oversight, AI
pations can also expect to benefit most tools could actually increase risks to
from the complementary benefits of AI— the financial system and undermine
think of a radiologist using the technol- financial stability.
ogy to improve her ability to analyze
medical images. For these reasons, the Public policy responses
overall impact in advanced economies Because AI operates across borders,
could be more polarized, with a large we urgently need a coordinated global
share of workers affected, but with only framework for developing it in a way
a fraction likely to reap the maximum that maximizes the enormous oppor-
productivity benefits. This has deeply disturbed schol- tunities of this technology while min-
Meanwhile, in emerging markets ars such as Yuval Harari. Through its imizing the obvious harms to society.
such as India, where agriculture plays mastery of language, Harari argues, AI That will require sound, smart poli-
a dominant role, less than 30 percent could form close relationships with peo- cies—balancing innovation and regu-
of employment is exposed to AI. Brazil ple, using “fake intimacy” to influence lation—that help ensure AI is used for
and South Africa are closer to 40 per- our opinions and worldviews. That has broad benefit.
cent. In these countries, the immediate the potential to destabilize societies. It Legislation proposed by the EU,
risk from AI may be reduced, but there may even undermine our basic under- which classifies AI by risk levels, is an
may also be fewer opportunities for standing of human civilization, given encouraging step forward. But globally,
AI-driven productivity boosts. that our cultural norms, from religion we are not on the same page. The EU’s
Over time, labor-saving AI could to nationhood, are based on accepted approach to AI differs from that of the
threaten developing economies that rely social narratives. US, whose approach differs from that of
heavily on labor-intensive sectors, espe- It’s telling that even the pioneers of the UK and China. If countries, or blocs
cially in services. Think of call centers in AI technology are wary of the existen- of countries, pursue their own regula-
India: tasks that have been offshored to tial risks it poses. Earlier this year, more tory approach or technology standards
emerging markets could be re-shored than 350 AI industry leaders signed for AI, it could slow the spread of the
to advanced economies and replaced a statement calling for global priority technology’s benefits while stoking
by AI. This could put developing econ- to be placed on mitigating the risk of dangerous rivalries among countries.
omies’ traditional competitive advan- “extinction” from AI. In doing so, they The last thing we want is for AI to
tage in the global market at risk and put the risk on par with pandemics and deepen fragmentation in an already
potentially make income convergence nuclear wars. divided world.
between them and advanced economies Already, AI is being used to comple- Fortunately, we do see progress.
more difficult. ment judgments traditionally made by Through the Group of Seven Hiroshima
humans. For example, the financial ser- AI Process, the US executive order on AI,
Redefining human vices industry has been quick to adapt and the UK AI Safety Summit, countries
Then there are, of course, the myriad this technology to a wide range of appli- have demonstrated a commitment to
ethical questions that AI raises. cations, including introducing it to help coordinated global action on AI, includ-
What’s remarkable about the latest conduct risk assessments and credit ing developing and—where needed—
wave of generative AI technology is underwriting and recommend invest- adopting international standards.
its ability to distill massive amounts of ments. But as another recent IMF paper Ultimately, we need to develop a set
knowledge into a convincing set of mes- shows, there are risks here. As we know, of global principles for the responsible
sages. AI doesn’t just think and learn herd mentality in the financial sector use of AI that can help harmonize legis-
fast—it now speaks like us, too. can drive stability risks, and a finan- lation and regulation at the local level.
DECEMBER 2023 71
F&D Straight Talk
72 D E C E M B E R 2023
Book Reviews F&D
Book Reviews
Inevitable
adequately integrate scientific knowl-
edge, and the profession in general for
failing to appreciate the severity of the
crisis and promoting a “persistently false
dichotomy” between sustainability and
growth. He advocates recognizing the
Simon Sharpe value of natural capital, thinking about
growth in terms of quality rather than
quantity, and revising the curricula of
raising the base of a house and mixing cement economics and business schools to teach
into its earth foundation make it less likely to collapse in a how growth can be made regenerative
flood. Backup generators and pumping systems reduce the instead of destructive.
risk of power and water supplies failing, allowing hospitals to If anything, this is one area the book
continue to operate. Robust health systems reduce the risk of could have pursued further. A shift in
disease following flooding. Education and communication thinking on the economics of decar-
systems can help people know when to stay home and when bonization has gathered pace in recent
to seek higher ground. Functioning public services increase years. Whereas once there was consen-
social trust, making it more likely that disaster response plans RISK AND sus that carbon pricing was the most effi-
will be implemented successfully. RESILIENCE cient solution, advances in the under-
IN THE ERA OF
In Risk and Resilience in the Era of Climate Change, Vinod standing of complex systems, studies of
CLIMATE CHANGE
Thomas makes a strong case for a systemic approach to build- technology transitions of the past, and
ing climate resilience. For starters, risk assessment should be Vinod Thomas observations of what is happening now
thorough and continuous. Investment is needed in early-warn- Palgrave Macmillan all suggest that approaches centered on
ing systems, evacuation plans, and institutions that govern all London, UK, 2023, investing in new solutions can be more
elements of disaster preparedness and response. Resilience 201 pp., $27.99 cost-effective in driving innovation and
must be integrated from the outset into the design not only structural change. It might have been
of buildings and infrastructure systems but also of national interesting to explore the application of
development strategies. this new understanding to the transfor-
Above all, Thomas emphasizes the need to prepare, not just mational adaptation the author argues
respond. Disaster proofing, he points out, adds less than a tenth is now needed.
to the cost of a new hospital. As climate change continues to Based on three decades of the author’s
progress, risks that once had low probability and high impact will work at the World Bank and the Asian
become high-probability, high-impact events. Governments Development Bank, and underpinned by
must anticipate this shift and continually prepare for the larger extensive academic research, the book
risks of the future, not merely for a repeat of events of the past. is full of practical case studies as well as
Thomas presents emissions reduction and adaptation as conceptual frameworks for understand-
intrinsically interdependent and goes so far as to describe ing resilience. The importance of these
decarbonization itself as an approach to building resilience. issues will only increase as the climate
Time is at the heart of this relationship: just as in the COVID- crisis progresses and extreme weather
19 pandemic, when slowing the spread of the disease was events inevitably become more com-
crucial to prevent hospitals’ intensive care units from being monplace; this will serve as a valuable
overwhelmed, slowing the onset of climate change will give guide to those working in the field. F&D
us more time to prepare for and cope with the extreme events
it will bring. simon sharpe is a senior fellow at
Perhaps the most bracing part of the book, for some readers, the World Resources Institute and author
is its chapter on economics. A distinguished economist himself, of Five Times Faster: Rethinking the
Thomas writes that “the mainstream economics profession Science, Economics, and Diplomacy
has not been on board in the campaign for climate action.” He of Climate Change.
DECEMBER 2023 73
F&D Book Reviews
74 D E C E M B E R 2023
Book Reviews F&D
technology alters the structure of inter- inequality is timely. As Walter Scheidel The erosion of modern values in
national comparative advantage. documents in his book The Great Leveler, advanced economies across recent
Visions of Inequality is an important which reviews the history of the con- decades has, according to Phelps, con-
scholarly work. But it is also a good read. sequences of inequality, large and per- tributed to the relative stagnation of pro-
Milanovic mixes his methodical exam- sistent increases in inequality can end ductivity and real wages. A task for eco-
ination of the evolution of economic up badly. F&D nomic policy, therefore, is to help society
thought about inequality with fascinat- regain these values in order to spark a
ing portraits of great economists and the zia qureshi is a senior fellow in new wave of dynamism and innovation.
society and polity of their times. the Global Economy and Development It will have to do so while dealing with
Renewed and deeper attention to program at the Brookings Institution. overwhelming challenges, including
climate change, digitalization, and the
plight of low-income earners.
DECEMBER 2023 75
F&D Currency Notes
On the Brink
Analisa R. Bala
The Philippines’ new 1,000 peso banknote was released last year.
with a wingspan of more than seven feet, steely protect cleared areas so that the birds can travel and hunt in
gray-blue eyes, and a mane-like crest, the Philippine eagle is safety. But if the eagles are to be saved, says Salvador, it will
one of the world’s largest and most striking raptors. It’s also at require a mass movement.
high risk of extinction, according to the International Union To raise awareness, two proclamations were issued by for-
for Conservation of Nature—a comprehensive information mer Presidents Fidel Ramos and Joseph Estrada: one declared
source of threatened animals, fungi, and plants. the eagle the national bird, the other established Philippine
Found on just four Philippine islands, the eagle is featured Eagle Week. In addition, a newly designed 1,000 peso poly-
on the new 1,000 peso banknote to “highlight the importance mer banknote was released in April last year, featuring the
of the preservation of this endangered species,” says Sara Cur- Philippine eagle and the national flower, sampaguita (a type
tis, a director at the Philippine central bank. But because the of jasmine), on the front. The eagle exemplifies “Filipinos’
birds are so difficult to track, there is still a “fundamental lack strength and love for freedom,” says Curtis.
of information regarding their distribution and population size,” The note’s design and security features won it the Interna-
Z U MA PR ES S, INC. /A L A MY STOCK P HOTO
says Dennis Salvador, executive director of the Philippine Eagle tional Banknote Society’s 2022 “Banknote of the Year” award,
Foundation (PEF) and coauthor of a recent study in the jour- and it is the first in the country to be printed on polymer.
nal Animal Conservation. Researchers estimate there are just Increased public awareness and a national wildlife con-
392 potential pairs left. servation law have bought the eagle time, but the PEF is
Eagle pairs need about 4,000–11,000 hectares of forest- breeding the birds in captivity to keep them from going
land to survive. Their range once covered 90 percent of the extinct. In an interview with Living Bird magazine, Salva-
Philippines, or about 27.5 million hectares of forest cover. That dor was hopeful: “I think we have a real chance of saving the
span has dwindled to only 7 million hectares, according to data eagles, even with the little we have left. It’s just a matter of
from the Department of Environment and Natural Resources. political will and attitude.” F&D
PEF—an institution dedicated to the survival of these
birds—is working with forest communities to reforest and analisa r. bala is on the staff of Finance & Development.
76 D E C E M B E R 2023
Timely.
Topical.
Free.