World Bank: Developing Countries Have Less to Fear From AI, but Depend on Outside Tech

World Bank's 2026 AI Report Makes the Case for Developing Economies

The World Bank's World Development Report 2026, The Promise of Artificial Intelligence, argues that low- and middle-income economies have less to fear from job automation and more to gain from AI than richer countries. The report estimates that less than one-tenth of jobs in developing economies are susceptible to AI automation, compared with more than one-third in high-income economies. At the same time, it puts 16.2% of developing-economy jobs in a more favourable category: roles that AI could complement rather than replace.

World Bank Group chief economist Indermit Gill writes that the technology could put expert knowledge within reach of millions rather than a small elite, and may compress what would otherwise take a century of economic transformation into a decade. Even under conservative assumptions, the Bank says, developing economies' potential growth rates would rise above the weak average of the first half of the 2020s.

The report's optimism comes with explicit warnings. Advanced AI models, the chips they rely on and the data centres that run them are controlled by a small number of companies in a few economies, creating dependency risks. The Bank notes that models such as ChatGPT are not trained on many of the more than 2,000 languages spoken in Africa. Physical infrastructure is another constraint: data centres could worsen water shortages and raise energy prices. At the 2025 Global AI Summit on Africa, 49 African countries and the African Union endorsed the Africa Declaration on Artificial Intelligence, stressing safeguards for human dignity, rights, freedoms and environmental sustainability.

What the World Bank's Data Shows About Jobs, Small Firms and AI Dependency

Why rich and developing economies face different AI job risks

The report's key distinction is about the structure of employment. Many developing countries have a high share of manual and small-firm work, while fewer people are in the knowledge-intensive office jobs where AI is most likely to replace tasks. That explains why the Bank estimates a much smaller share of jobs are exposed to automation in low- and middle-income countries than in high-income ones. The immediate disruption risk is concentrated in knowledge-intensive professional services, not in the informal and manual occupations that dominate many developing labour markets.

Small businesses may benefit before large employers do

Because nearly 90% of workers in lower-middle-income countries are employed in firms with fewer than ten people, the report sees AI as a tool that can reach small businesses through the apps they already use, such as messaging platforms, business software and social media. This is not a future hypothetical: the Bank cites Kenya's judiciary, which used AI to assign more than 10,000 court cases a year to over 1,500 mediators, helping to reduce a large backlog. Similarly, teachers in Sierra Leone have used AI to tailor syllabi, and small businesses in Kenya have accessed AI advice through WhatsApp.

Dependence on foreign technology is the main brake on local gains

The report's most sobering point is structural: the most advanced models, chips and data centres sit with a few companies in a few economies. That limits whether AI tools meet developing countries' needs. The language gap is one symptom; the infrastructure bill is another, because data centres may compete with households for water and raise energy costs. Without reliable electricity, internet access and strong education systems, the report says, developing countries cannot fully capture AI's benefits.

African governments are already moving on governance

The Africa Declaration on Artificial Intelligence, endorsed by 49 African countries and the African Union, shows that governments are not waiting for foreign providers to set all the rules. The report frames local adaptation, not simply importing ChatGPT-style models, as the largest opportunity. That means small AI solutions built for specific development challenges and existing resources, rather than only large generic systems.

Priority Moves for Governments and Small Businesses in Developing Countries

For governments and development institutions, the report points to a clear sequence of practical investments and rules.

• Prioritise reliable electricity, internet access and basic education before scaling AI, because the World Bank identifies these as prerequisites for capturing AI gains in developing economies.
• Design small-business programmes around AI features already embedded in messaging apps, business software and social media, since nearly 90% of lower-middle-income workers are in firms with fewer than ten people and more than half are self-employed.
• Replicate proven public-sector uses such as Kenya's court mediation model, where AI assigned more than 10,000 cases a year to over 1,500 mediators to reduce case backlogs.
• Convert the 2025 Africa Declaration on AI into concrete procurement and data rules, because dependency on a small number of foreign AI providers can make available tools poorly suited to local needs.
• Fund foundational skills, including literacy, numeracy, reasoning, basic digital skills and teamwork, because the report says these determine whether workers benefit from AI.