Artificial intelligence will have a relatively small immediate effect on Ethiopian jobs, but weak digital infrastructure, skills shortages and limited local data could hold back the productivity gains the technology offers, according to the World Bank’s World Development Report 2026.
The assessment carries an implicit challenge to Ethiopia’s stated ambitions. Prime Minister Abiy Ahmed has positioned the country as a leader in Africa’s technological transformation and an advocate for digital sovereignty, pointing to the Ethiopian Artificial Intelligence Institute, the Digital ID programme and the 5 Million Coders campaign. Deputy Prime Minister Temesgen Tiruneh described the launch of a university AI platform in February as marking Ethiopia’s transition “from technology consumer to technology producer.”
The World Bank’s framework says that sequence should run the other way.
Adopt, adapt, advance
The report structures its recommendations in three stages. For most developing economies, it argues, adopting existing AI applications and adapting them to local conditions offers a more practical route than building frontier systems, which requires substantial investment in computing infrastructure, data centres, data and highly specialised talent.
The Bank recommends developing countries first strengthen the infrastructure and skills needed to use existing AI productively, before committing large sums to frontier development.
For Ethiopia, the report says, adaptation matters particularly because systems built for other markets may perform poorly against local languages, institutions, agricultural conditions and public services. It cites African companies building AI models for African languages as the model to follow.
Low exposure, modest gains
The Bank groups Ethiopia with Vietnam as economies where employment concentrates in agriculture and manufacturing — sectors with relatively low exposure to AI-driven automation. The immediate employment effect is expected to be modest, and so are the productivity gains.
It estimates that 4.5% of existing jobs in low- and middle-income economies are potentially exposed to automation by generative AI, against 14.2% in high-income economies. At the same time, 16.2% of jobs in lower-income economies could be complemented rather than replaced by AI.
That framing echoes the International Labour Organization’s recent finding, covered by iAfrica this month, that AI displacement risk concentrates in wealthier countries while the binding constraint in developing economies is informality rather than automation.
The infrastructure numbers
The report’s constraint data is stark. About three-quarters of firms in sub-Saharan Africa report experiencing power outages. Internet speeds in low- and lower-middle-income countries generally sit below 25 megabits per second, against more than 300 Mbps in high-income countries.
Outside cities the gaps widen. In 2024, 32% of rural schools in sub-Saharan Africa lacked regular electricity and 68% lacked consistent internet access. The report also notes that 89% of 10-year-olds in the region cannot read and understand simple text — a foundational skills problem that sits beneath any digital skills agenda.
Those conditions limit what businesses, farmers, schools and government agencies can do with AI, particularly where applications depend on continuous connectivity. The Bank says applications need to work on basic devices and weak networks to reach populations outside well-connected urban centres.
Microsoft’s own measurement supports the caution. Its Global AI Diffusion Q1 2026 report placed Ethiopia below 8% generative AI adoption among the working-age population — near the bottom of the 147 economies it measured, and far behind South Africa’s 23.1%.
Where Ethiopia does stand out
Ethiopia has a national AI policy and appears on the World Bank’s map of countries with a national AI strategy — a list that excludes South Africa, whose draft policy was withdrawn in April over fabricated citations. More than 80 countries had published strategies by June 2026, though the Bank notes that only one of 25 low-income countries, Rwanda, had a dedicated strategy.
AI also sits within Digital Ethiopia 2030, the country’s broader digital transformation programme covering infrastructure, skills and technology-enabled public services, and Ethiopia has pursued international AI cooperation including with India on agriculture and health applications.
What the Bank is actually asking for
The report’s conclusion is that Ethiopia’s near-term challenge is not building a rival to the world’s largest AI companies. It is creating conditions in which businesses and government agencies can use existing AI productively: expanding electricity and broadband, improving digital and foundational skills, developing local datasets, strengthening data governance, and enabling local companies to adapt applications to domestic needs.
Countries that fail to build those foundations, the Bank warns, risk watching productivity gaps widen even as AI adoption accelerates elsewhere.
That is not incompatible with Ethiopia’s sovereignty argument — building local datasets and adapting applications to Amharic, Oromo and Tigrinya is itself a form of technological self-determination. But it is a different proposition from continental AI leadership, and it starts from a lower rung.
