Artificial intelligence (AI) is driving trillions of dollars of investment, particularly in the United States. But despite growing adoption, there is little evidence so far of major global impact on productivity and labour markets. As AI progresses towards becoming a general-purpose technology with global macroeconomic impacts, uncertainties remain over their magnitude, pace and geographic distribution.
Artificial intelligence is moving towards becoming an established general-purpose technology, with the potential to transform sectors, boost productivity and reshape growth prospects. However, its benefits will not be evenly distributed. The international dimension of this transformation will depend not only on who develops the most advanced technologies, but also on each economy’s ability to adopt them and integrate them into its production processes.
“AI adoption is accelerating, especially among large firms and knowledge-intensive sectors, but its diffusion remains uneven across companies, industries and countries”
The United States and China hold central positions in AI development, with different strengths in talent, research, investment and infrastructure. Yet technological competition alone does not determine how economic gains are distributed. Countries that do not lead in model development can reap significant benefits if they succeed in broadening adoption and translating it into tangible productivity improvements.
Business adoption is accelerating, although it remains uneven across countries, sectors and firm sizes. There is evidence of greater use in the United States than in Europe, and of adoption being concentrated among large companies and knowledge-intensive activities. These differences reflect the composition of the business sector, organisational capabilities and the conditions that facilitate the integration of new tools.
“AI leadership is becoming a strategic race: countries that dominate talent, research, infrastructure and models may capture both economic gains and geopolitical influence”
In this context, each country’s preparedness is crucial. Digital infrastructure, human capital, innovation capacity and the regulatory framework shape the speed and depth of adoption. Economic structure also plays a role: financial, technological and scientific activities, along with other knowledge-intensive services, have greater exposure to AI. Labour costs, in turn, affect incentives to automate tasks or complement human work.
Using AI, however, does not automatically translate into higher productivity across the economy. Although numerous studies find improvements in specific tasks, the aggregate effect remains difficult to identify. Time savings can be eroded by bottlenecks, coordination costs or subsequent stages of review and integration. Turning these improvements into real output requires organisational changes that go beyond adopting a technological tool.
This gap between technical capability and economic outcomes helps explain why the transformation is gradual. As with other major innovations, productivity gains may take years to materialise.
The estimates presented in the article point to greater potential benefits in advanced economies, albeit with significant differences within each region. Some small, digitally advanced, knowledge-specialised economies could achieve gains comparable to those of the technological leaders. By contrast, insufficient preparedness could limit the benefits in many emerging and low-income economies.
The key international question is not simply who will lead in AI, but who will succeed in spreading its use and making the most of it. AI’s impact will depend on how each country translates technological potential into productivity and growth.