How AI could boost global growth — and widen the divide between rich and developing economies

The warning comes amid a surge in investment in semiconductors and AI infrastructure, which has helped support global economic growth and propel stock markets to record highs. Yet the scale of spending has also raised questions about whether companies can generate adequate returns from their investments, particularly as borrowing costs rise and evidence of economy-wide productivity gains remains limited.

Speaking at Reuters Global Markets Forum, IMF and OECD officials highlighted a challenge facing the global economy: investment in artificial intelligence is accelerating, but its adoption across businesses and countries remains uneven. For developing economies, the ability to integrate AI into business operations could determine whether they benefit from the technology-led growth cycle or fall further behind countries with better infrastructure and greater access to capital.

5% in Q3: What drove growth and what remains a concern AI could add to global growth, but productivity gains remain limited The IMF estimates that effective adoption of artificial intelligence could increase annual global economic growth by up to 0.5% over the medium term. However, the potential gains will depend on whether countries can move beyond access to the technology and deploy it widely across their economies.

"For many developing economies, the critical question is whether they can adopt and diffuse AI," Era Dabla-Norris, deputy director of the IMF's Fiscal Affairs Department, told the Reuters Global Markets Forum. Dabla-Norris cautioned that the productivity benefits visible across the wider economy have so far been less pronounced than the scale of investment might suggest. "But the aggregate productivity evidence, so far, is much less," she said.

The OECD estimates that AI could add between 0.4% and 1.2% to annual productivity growth across its member countries. However, OECD chief economist Stefano Scarpetta said the technology's adoption in workplaces was lagging behind its use by consumers, with businesses making limited use of AI in their core functions. "There is a large gap between exposure and actual use," Reuters quoted Scarpetta as saying.

The gap matters as the economic value of AI depends on how effectively businesses incorporate it into production, services and decision-making. Investment in computing infrastructure alone may not translate into higher productivity unless companies can reorganise workflows, develop relevant skills and use the technology at scale. Reliable electricity, digital connectivity and access to skilled workers are essential to deploying AI across businesses, particularly beyond large technology companies.

Developing economies could struggle to capture the potential gains if these prerequisites remain unevenly distributed. Smaller companies may also find it harder to invest in the infrastructure, tools and expertise required to integrate AI into their operations. This creates the risk of a widening productivity divide. Businesses with access to capital and advanced technology could use AI to reduce costs and improve efficiency, while firms with fewer resources may be unable to adopt it at the same pace.

The IMF's growth estimate represents a potential economic gain rather than a guaranteed outcome. The extent to which countries benefit will depend on how widely AI is adopted and whether businesses and workers have the capacity to use it effectively. AI investment boom raises questions over returns The rapid expansion of AI infrastructure has become an important source of investment, particularly in chips and the computing capacity needed to develop and run AI systems.

That spending has helped support global growth and stock market gains. However, companies are under pressure to demonstrate that the capital committed to AI can generate sustainable financial returns. Higher borrowing costs add to that challenge by increasing the cost of financing large investment programmes, Reuters reported. For businesses, the key question is whether AI can deliver measurable gains through lower operating expenses, higher output or additional revenue.

The limited evidence of broad-based productivity improvements so far suggests that the economic payoff may take time to emerge. The risk is not necessarily that AI investment will fail, but that the gains may arrive more slowly or be concentrated among a smaller group of companies and economies than investors expect. 3%: IMF warns of sharp downturn for Pakistan, MENA AI could reshape employment, with entry-level workers vulnerable The economic consequences of AI extend beyond productivity and investment to the labour market.

The IMF estimates that around 40% of jobs worldwide could be exposed to the technology, according to Reuters, as companies increasingly explore automation and other digital tools to reduce costs. Recent workforce reduction announcements have added to concerns about the potential impact on employment. Barclays said it was cutting around 5,000 jobs this year as it uses technology, including AI, to reduce costs and improve returns.

Reuters also reported planned job cuts at eBay and Qualcomm. However, exposure to AI does not automatically mean that a job will disappear. The effects will depend on the tasks involved, how employers deploy the technology and whether workers can adapt to changing requirements. Dabla-Norris noted a particular risk for younger and entry-level employees. If AI takes over routine assignments through which workers traditionally gain experience, new entrants could lose opportunities to develop skills on the job.