AI Investment Supports Growth but Puts Young Workers and Economies at Risk
Table of Contents
You might want to know
- How can artificial intelligence investment support economic growth while also increasing uncertainty for workers and governments?
- Why might young people entering the workforce experience AI’s effects sooner than other workers?
Main Topic
Artificial intelligence investment is helping sustain global economic growth even as an energy shock weighs on many countries, according to International Monetary Fund (IMF) Managing Director Kristalina Georgieva. Speaking Monday in Bangkok on the opening day of the IMF and World Bank annual meetings, she said the world economy is growing at roughly 3% a year. The meetings run through October 18 in the Thai capital.
Georgieva described AI as a “slow-moving tsunami”: a technology whose effects are building over time and are expected to reach a substantial portion of the global workforce. The IMF expects AI eventually to affect about 40% of jobs worldwide, including roles that may be changed, downgraded, or eliminated. The scale of exposure differs significantly by economy, with around 60% of jobs in advanced economies exposed, compared with roughly 23% in low-income countries.
Those figures do not mean that every exposed job will disappear. Exposure refers to the potential for AI to alter tasks or how work is performed. In some roles, software may take over particular activities; in others, it may support workers or change the skills employers require. Georgieva noted that lower exposure in poorer countries is partly because AI has not yet reached them to the same extent. That gap may therefore reflect differences in access and adoption, not immunity from future change.
Young people starting their careers may be especially vulnerable because entry-level positions often depend on routine tasks that AI systems can increasingly perform. Georgieva said that new workers are first in line for this disruption. Singapore President Tharman Shanmugaratnam, who spoke on the same panel, said demand for graduates’ first jobs is already falling. These concerns point to a potential change in the way employers recruit and train junior staff, even where AI does not replace an entire occupation.
The economic effects of AI are not limited to employment. Georgieva said AI investment is helping keep overall growth near 3% despite rising energy costs. In her account, however, the gains are concentrated in places that manufacture the hardware and related products behind the technology. Asian economies that supply AI infrastructure could receive a particularly large boost as adoption expands.
According to Georgieva, AI adoption could add 0.5 to 1 percentage point a year to growth in advanced economies and 0.2 to 0.8 points in emerging economies. A percentage point is the difference between two growth rates: for example, an economy growing 2% that gains a full percentage point would grow 3%. Georgieva said the high end of the projected range could more than offset the drag from aging populations across the region.
The IMF’s July outlook update found that the top four net exporters of AI hardware—Taiwan, South Korea, Thailand, and Malaysia—outperformed the Fund’s first-quarter growth forecasts by an average of 4.4% on an annualized basis. South Korea grew at a 7.5% pace, compared with the 1.8% the IMF had projected in April. These comparisons illustrate how the expansion of AI-related manufacturing can benefit economies positioned within the supply chain, although they do not establish that AI investment alone caused the growth differences.
Georgieva also said AI hardware and related technology products now account for more than a tenth of world goods trade. She added that AI investment as a share of global gross domestic product will likely surpass the investment that went into building railroads or the electricity grid. The comparison conveys the potential scale of the build-out, while also highlighting the substantial resources being committed to computing infrastructure and related technologies.
At the same time, many countries remain outside the main AI supply chain and may not share in the expansion to the same degree. The IMF’s July update projected 2026 global growth at 3%, down from 3.5% in 2025, as the Middle East war drove energy prices up. AI investment may help counter some economic pressures, but the projected gains do not remove the effects of higher energy costs or guarantee that growth will be distributed evenly.
Uneven access could also widen the divide between richer and developing economies. Shanmugaratnam warned that AI could increase this gap if governments do not respond. Differences in connectivity, computing resources, investment, and skills may affect whether businesses and public services in lower-income countries can use AI effectively. The issue is therefore not only how quickly the technology advances, but also who can obtain and apply it.
Public finances are another area of concern. Georgieva said that if AI shifts income away from workers and toward owners of capital, governments could lose part of the tax base on which they rely. The consequences would depend on how income and employment evolve, as well as on tax systems and policy choices. Still, a substantial shift in the distribution of earnings could complicate efforts to fund public services and support workers facing changes in their occupations.
World Bank President Ajay Banga described one effort to address unequal access: backing “small AI” tools that do not depend on large language models. These tools are aimed at practical uses in emerging markets, including applications for farmers, clinics, and classrooms. Such approaches may be useful where large-scale computing infrastructure is limited, though their effectiveness will depend on local needs, reliable access, and appropriate oversight.
The rapid expansion of AI also presents financial and inflation-related risks. Georgieva said in Singapore that the construction and deployment of AI systems are adding to inflation. She also warned that disappointing earnings could spread through markets because major AI companies have borrowed heavily to fund their build-out. A reversal in expectations could therefore affect not only technology firms but also lenders and investors connected to the sector.
The IMF’s full World Economic Outlook was scheduled for release at 9 a.m. Bangkok time on October 13, corresponding to 10 p.m. Monday in New York. It was set to provide the Fund’s first revision to its 3% global growth forecast since July. That publication would offer a broader assessment of how the Fund views the interaction between AI investment, energy pressures, and economic prospects.
Key Insights Table
| Aspect | Description |
|---|---|
| Global growth and AI investment | Georgieva said AI investment is helping keep global growth near 3% a year despite an energy shock. |
| Jobs affected | The IMF expects AI eventually to touch about 40% of jobs worldwide, around 60% in advanced economies, and roughly 23% in low-income countries. |
| Young workers | Entry-level roles often include routine tasks that AI can absorb, potentially reducing opportunities for young people and graduates entering the workforce. |
| Growth opportunities in Asia | AI adoption could add 0.5 to 1 percentage point a year to growth in advanced economies and 0.2 to 0.8 points in emerging economies. Taiwan, South Korea, Thailand, and Malaysia are major net exporters of AI hardware. |
| Distribution and public finances | Benefits may bypass countries outside the AI supply chain, while a shift in income from workers to capital owners could reduce part of the tax base governments depend on. |
| Investment risks | Georgieva said the AI build-out is adding to inflation and cautioned that disappointing earnings could affect markets, as major AI companies have borrowed heavily to finance expansion. |
| IMF outlook | The IMF projected 2026 global growth at 3%, down from 3.5% in 2025, and scheduled its full World Economic Outlook for October 13. |
Afterwards...
AI’s economic effects will depend on more than the technology itself. Governments, employers, educators, and researchers will need to understand how tasks change across occupations and whether workers can move into new roles. Better evidence on hiring, wages, productivity, and job transitions can help distinguish between temporary changes and lasting shifts in employment.
Education and training systems may need to give greater attention to the skills that complement automated tools, including critical judgment, communication, technical understanding, and the ability to work with AI responsibly. Particular focus on early-career pathways could help ensure that young workers still have opportunities to learn through experience when routine assignments become automated. Employers can also examine whether they are redesigning entry-level jobs and training, rather than simply removing junior positions.
Further work is also needed to make AI accessible beyond the economies that produce its hardware. Research into lower-cost systems, locally relevant tools, dependable connectivity, and responsible data practices could help expand practical benefits in emerging markets. Initiatives such as “small AI” merit careful assessment in settings like agriculture, health care, and education, where usefulness, reliability, and local control are especially important.
Finally, policymakers and financial institutions will need to monitor the distribution of AI-driven income, its effects on public revenue, and the financial exposure created by borrowing to fund infrastructure. The central challenge is to turn AI’s investment boom into broad-based productivity and opportunity while preparing workers and public institutions for disruption. Continued study of these areas can help societies respond to the technology’s benefits and risks without assuming that growth alone will resolve unequal outcomes.
Last edited at：2026/10/12