AI could transform some of the world's largest and most labor-intensive industries. Yet its application and adoption, both across and within nations, will likely be uneven. The differences that emerge could affect global growth and the nature of employment, including through the creation of new jobs, improvement to some workers' efficiency, and by replacing some roles.

AI will compete for investment with three other global, megatrends: the sustainable energy transition, digitization, and the aging population (see "Assessing How Megatrends May Influence Credit Ratings," April 18. 2024). Isolating the effects of AI-fueled disruption from the economic effects of that trio will be difficult given the widespread, multifaceted, and often overlapping pressures and opportunities that will emerge.

S&P Global Ratings expects four factors will determine which economies benefit most from AI:

AI Development Will Be Multispeed And Multidimensional

Wealthier economies could benefit from an initial advantage in terms of AI development and deployment due to existing strengths, notably in public and private investment capacity, digital infrastructure, and workforce capability.

Many of the major developments in technologies that incorporate AI are taking place at large, publicly traded companies and the integration of AI into industrial applications is a logical extension of that. Wealthier economies with deep domestic capital markets and more developed Information and Communication Technology (ICT) sectors will have greater capacity and scale to absorb the short- to medium-term increase in investment that will be required to foster AI-development.

We expect AI will be economically disruptive before potentially emerging as a multiplier that provides advantages to first movers. However, the extent and timing of that impact is unclear, for now. For example, the investment demands relating to AI's application in the public sector might initially lead governments to defer spending on other projects, but the technology should ultimately deliver savings and efficiencies that create new scope for future investment. Disruption is also likely in labor markets, where AI will create redundancies alongside new jobs, and over the longer-term foster skills migration and upskilling (see "Can generative AI create a productivity boom?" Jan. 10, 2024).

In the long run, it is uncertain to what extent the effective deployment of AI will enhance economic growth. That is notably because of doubts about the extent to which AI-related productivity gains could be offset by rising unemployment. At the same time, the ability to measure (and isolate) the extent of economic growth attributable to AI will be complicated by changes in demand for products and services and due to externalities, such as global mega trends or crises.

AI's deployment will come with risks. For example, rapid implementation across numerous sectors (and without sufficient testing, security, or safeguards) could result in unexpected financial costs, operational risks, and increase systems' vulnerability to cyberattacks. Malevolent application of AI could also foster more sophisticated and unpredictable espionage capabilities, potentially with severe consequences for industry and political stability.

State Versus Private: AI Funding Models Will Differ Across Nations

The extent to which some countries are already establishing an early AI lead is evidenced by the differences in private-sector investment in the technology. The U.S., China, and the U.K. accounted for about 81% of global investment spending on global AI startups over the period from 2013-2023 (see chart 1), equal to about 0.3% of those countries' aggregated GDP over the same period.

The U.S.'s current dominance of this metric reflects an economy driven by the private sector as well as institutional strengths--including sound policy support, vast financial resources, a wide talent pool, and the existence of large, privately led innovation hubs. This is also evident in the creation of about 5,500 new AI-related companies in the U.S. since 2013, which is more than all the other countries combined and over 3.6 times more than second-placed China (see chart 2).

Much of China's spending on AI is likely to be directly government funded and could be relatively opaque to outside observers. For this reason, comparisons of AI investment spending among nations should be treated with caution.

AI Investment And Activity Will Accelerate

Over the next three years, we expect the creation of new AI companies will accelerate along with the pace of investment. We anticipate global private investments (i.e., excluding public sector funding and spending by established larger corporates) in AI startups could reach $800 billion to $900 billion by 2027. This implies a compound annual growth rate (CAGR) of at least 70% to 74%, which assumes growth of private investment in AI at 1.0x to 1.5x of forecast revenues from generative AI companies for the respective regions, per annum until 2027 (see chart 3a and chart 3b, and "Generative AI Market Monitor & Forecast," June, 27, 2024, published by 451 Research).

Our expectation of a significant uptick in AI investment from 2023 (as reflected in chart 3a and chart 3b) stems from the spending increases that are accompanying the deployment of generative AI. The benefits of deployment have notably prompted a revision of earlier, national AI-strategies, many of which were developed and published from 2017 to 2018. Some key, recent AI investment initiatives announced by selected governments are summarized below (see table 1).

The U.S. Department of Commerce's CHIPS Act is a recent example of the sort of financial incentive that could bolster semiconductor and research capabilities in the U.S. (see "CHIPS Act Funding To Intel Sparks Revival For U.S. Semiconductor Manufacturing," March, 27, 2024). Similarly, the European Union is working to boost EU competitiveness, focusing on research and industrial capacity, particularly for startups and small and medium-sized enterprises (SMEs).

We consider that geopolitical uncertainties related to the Russia-Ukraine war or the Israel-Hamas conflict could also lead to reallocation of AI spending towards defense spending in some countries. That said, the reallocation could also result in an increase in military spending on AI-related projects.

The Importance Of Digital Infrastructure And AI Policies

While both private and public sector investment will be central to determining AI's impact, we also think that potential near-term gains from AI (design, deployment, and usage) will accrue more readily in economies with stronger digital infrastructure--including computing capacity, broadband speed, and data productivity. Additionally, suitable policy frameworks, including regulation and governance, will be essential to facilitating AI's application and the resulting economic gains.

Quantifying the benefits of strong infrastructure and supportive regulatory and governance factors is difficult, not least due to the shifting requirements of the rapidly evolving AI-landscape. Yet it is apparent that some nations are better positioned than others, at least for now. The U.S, Singapore, the U.K., Finland, and Canada all benefit from AI-friendly policies, regulations, and monitoring systems as well as competitive technology sectors, data availability and quality, and strong infrastructure, according to the "AI Readiness Index," produced by Oxford Insights (see chart 4). We note that there are numerous rankings of national AI preparedness and deployment, such as the "Global AI Index" from Tortoise Media, among others.

In terms of regulation, we expect jurisdictions that have developed human-centric, risk-based, and balanced AI regulations will be best placed to control the technologies' impact, including on labor markets and inequality. The EU has established itself as a leader in this respect, following the approval of the EU AI Act, which we expect will influence companies and regulators across the globe (see "Your Three Minutes In AI: The EU AI Act Could Become A Global Benchmark," March 15, 2024). Another example of the emerging regulatory environment is the "Seoul Declaration for safe, innovative and inclusive AI," which was signed by 10 countries and the EU, in May 2024. It lays the foundations for an international network of publicly-backed AI-Safety Institutes, establishes a common understanding of AI safety, and seeks to create a framework for AI research based on trust and responsibility. The Declaration could ultimately prove the basis for widely adopted, best-practice protocols for AI-safety.

Labor Is An Important Factor

Despite fears that AI will replace humans, the advancement of the technology remains dependent on access to a qualified labor force. The number of AI-related jobs (and postings) has grown steadily in recent years. We expect that trend will continue through 2024 and the years to come, supported by AI regulation and a shift from AI testing to deployment at scale. Unsurprisingly, the U.S. is the largest job market for AI-related roles, accounting for about 1.6% of postings in 2023, followed by Spain on 1.4%, and Sweden 1.3%, (see chart 5). While Spain and Sweden do not appear at the top of AI-related investment spending, job openings could be a leading indicator for upcoming AI-related investments.

AI's Economic Impact Remains Uncertain

In theory, a rough measure of the economic impact of AI could be provided by estimating AI-led labor productivity gains (real GDP per hour worked). The potential for AI to increase productivity promises to be novel, at first, and may yet prove uncertain. It will also be difficult to measure and detangle AI's effects on productivity from other major global disruptions, including the energy transition, digitization, and the effects of an ageing population in many nations. Moreover, it is notable that earlier transformative technologies, such as the personal computer and widespread access to the Internet, did not necessarily result in productivity improvements visible in economic statistics--a phenomenon that came to be known as Solow's Paradox.

Ongoing gains from AI will rest on continued and significant investment from the private and public sector, widespread adoption of the technology, continued improvement of digital infrastructure, proactive and effective policies and regulation, and an adequately skilled and sized labor force. Even if all those factors are in place, there could still be economic downsides from AI development, such as the disincentivizing of innovation due to AI facilitating the replication of advances.

The labor market for AI-linked jobs is itself exposed to disruption. Demand for skilled AI-workers could fuel migration, encouraging countries to offer incentives to attract skilled labor. This competition, and the resulting advantage for winners, could lead to shifts in economic gains and the wider employment outlook for nations.

What To Watch Over The Next Five-to-10 Years

While AI is often referred to as a revolution, we expect its economic impact will be felt only gradually. In the short-term those effects may be negatively weighted, though we expect greater balance over the longer-run as economic benefits emerge (see chart 6). More specifically, we expect that effective deployment of AI will reduce costs through increased efficiency and thus create revenues from existing and new sources that had been too costly to develop. The upshot of that should be positive economic growth.

Those gains will be balanced against risks. Most notably (in economic terms), AI will first weigh on private and public sector investment spending, may in some instances have uncertain outcomes, and will likely take time to yield returns. And that investment will happen against a backdrop of other factors that also affect economies, including pressures at varying points in the economic cycle and unpredictable crises.

Labor market spillovers:  The tradeoff between job displacement, from AI-powered automation, and the improvements resulting from AI-enhanced labor productivity is a complex issue. Younger generations should reskill over the next five years, but there's a risk that the wider pool of human talent takes longer to adapt to AI-induced changes. And there is the risk that highly skilled workers will migrate quickly to economies that emerge as AI leaders (and thus have an attractive job market). These risks could exacerbate inequalities between and within countries and create a permanent competitive disadvantage for economies that experience a talent deficit due to migration.

Regulatory-induced competitiveness:  We're already observing different regulatory approaches to AI. Requirements for companies to comply with stricter regulations in some regions will lead to slower AI innovation, particularly for large foundation models (i.e., AI models with diverse applications that are trained on vast quantities of usually unlabeled data) that are more complex and carry a higher degree of unforeseen risks. Regulatory treatments could also inhibit AI application in certain economies, potentially leading to competitive disadvantages compared to more liberal jurisdictions.

Operational resilience:  Nations will feel pressure not to be left behind in the AI-adoption race. However, this fear-of-missing-out needs to be balanced against strategic planning, effective AI governance, and a human-centric approach. Entities that rush to deploy AI without careful consideration of those factors will be more prone to unintended consequences, including the potential for operational, reputational, and financial damage.

Related Research

• Assessing How Megatrends May Influence Credit Ratings, April 18, 2024
• CHIPS Act Funding To Intel Sparks Revival For U.S. Semiconductor Manufacturing, March, 27, 2024
• Your Three Minutes In AI: The EU AI Act Could Become A Global Benchmark, March 15, 2024
• Can generative AI create a productivity boom? Jan. 10, 2024

Writer: Paul Whitfield