- Developing Economies Could Harness AI Without Building Costly Frontier Models — World Bank
Artificial intelligence could give developing economies an unusual opportunity to compress decades of conventional development into a much shorter period, offering a new source of productivity growth at a time when poorer countries are experiencing their weakest average economic performance in three decades, according to the World Bank.
The World Bank Group’s World Development Report 2026: The Promise of Artificial Intelligence argues that countries do not necessarily need to build frontier AI models or construct vast data centres to benefit from the technology. Instead, relatively inexpensive AI applications adapted to local conditions could improve healthcare, education, agriculture, government administration and access to essential services.
“AI has thrown developing economies a lifeline, and they should seize it,” said Indermit Gill, Senior Vice President and Chief Economist of the World Bank Group. “They do not need large models or big data centers to reap its benefits. By adapting small, low-cost AI tools to local conditions, they can bring better medical care, education, judicial services and agricultural extension within reach of millions.”
The report, however, presents AI neither as an automatic growth engine nor as a substitute for traditional development. Its impact will depend heavily on whether countries can provide reliable electricity, affordable internet connectivity, digital skills, computing capacity, locally relevant data and institutions capable of governing increasingly powerful technologies.
The employment implications are particularly striking. The World Bank estimates that 14.20% of jobs in high-income economies face potential automation from generative AI, more than three times the 4.50% estimated for developing economies, reflecting the greater concentration of clerical, professional and knowledge-intensive jobs in richer countries.
For developing economies, the potentially bigger story is augmentation rather than replacement. About 16.20% of jobs could experience meaningful productivity improvements through AI, compared with 18.70% in high-income economies, suggesting poorer countries may be able to capture substantial productivity gains without immediately facing the same scale of labour displacement.
That could be particularly important where shortages of skilled workers already constrain development. AI-assisted diagnostic systems could extend the capabilities of health workers, teachers could adapt educational materials more efficiently, farmers could receive location-specific information on weather and crop diseases, while small businesses could gain analytical capabilities previously available mainly to larger companies.
Governments could also deploy AI to strengthen tax administration, target social programmes, improve disaster response and make public services more accessible. For countries constrained by limited fiscal space and shortages of specialist personnel, productivity improvements of that nature could allow existing resources to stretch further.
The opportunity arrives as traditional growth engines are becoming harder to rely upon. Elevated public debt, constrained government budgets and a more fragmented global trading environment have complicated the development strategies available to emerging and low-income economies, increasing the attraction of technology-driven productivity gains.
The World Bank consequently proposes an “adopt, adapt and advance” strategy. Countries should initially deploy proven applications, adapt them to domestic languages and economic conditions and only move towards more sophisticated AI development as their technological and institutional capabilities deepen.
That sequencing is particularly relevant for Africa because it would allow governments to avoid spending scarce resources trying to replicate the infrastructure of global technology leaders. The objective would instead be to solve specific domestic problems with technologies that already exist.
But the report also exposes a fundamental contradiction: the world’s newest general-purpose technology remains dependent on some of development’s oldest infrastructure. In Sub-Saharan Africa, nearly one-third of rural schools lack reliable electricity and more than two-thirds lack dependable internet access, potentially excluding millions from the productivity gains AI promises.
That makes electricity and broadband policy inseparable from AI strategy. The World Bank Group’s Mission 300 initiative, which aims to provide electricity access to 300 million people in Sub-Saharan Africa by 2030, increasingly represents not only an energy programme but a gateway into the digital economy.
For economies such as Ghana, the implication is clear. Ambitions to become AI-enabled will depend on continued investment in electricity reliability, fibre networks, affordable mobile data, cloud infrastructure, digital education and cybersecurity rather than simply publishing national AI strategies.
Data will become another important economic asset. AI systems trained primarily on information generated in advanced economies may struggle with African languages, local farming conditions, legal systems and social contexts, increasing the importance of building high-quality domestic datasets.
The World Bank also cautions against confusing experimentation with progress. Governments may launch dozens of AI pilots, but their economic value will depend on whether they measurably improve productivity, educational outcomes, healthcare delivery or administrative efficiency.
“The window to get this right is narrow,” said Gaurav Nayyar, Director of the World Development Report 2026. “AI presents a once-in-a-lifetime opportunity to solve problems that have resisted solutions for generations.”
There are equally significant risks. Frontier AI capabilities remain concentrated among a relatively small number of companies and countries, meaning technological diffusion could either narrow development gaps or deepen them, particularly if poorer economies remain dependent on systems, infrastructure and intellectual property controlled elsewhere.
Similar inequality could emerge domestically between connected urban populations and underserved rural communities, large firms and smaller enterprises, or workers possessing advanced digital skills and those unable to participate in the new economy. Privacy, algorithmic bias, misinformation and accountability will consequently require increasing regulatory attention.
The World Bank’s central proposition is therefore less about whether AI is powerful enough to transform development than whether developing countries are institutionally and economically ready to use it. They do not have to win the global contest to build the largest models, but they must create the foundations on which useful applications can operate.
If they succeed, AI could become a productivity multiplier capable of allowing countries to bypass some of the constraints that historically slowed development. If they fail, a technology capable of narrowing decades of lost progress could instead create another divide between economies able to harness the AI revolution and those left consuming technologies designed elsewhere.
