World Bank Puts Developing-Economy AI Automation Risk at 4.5 Percent of Jobs
Only 4.5 percent of jobs in developing economies face automation risk from generative artificial intelligence, against 14.2 percent in high-income countries, the World Bank Group said on 4 August in its World Development Report 2026, subtitled The Promise of Artificial Intelligence. Jobs in wealthy nations are, in the bank’s phrasing, more than three times as likely to face that risk.
The report sets a second pair of figures against the first. Some 16.2 percent of jobs in developing economies could see their productivity meaningfully boosted by artificial intelligence, close to the 18.7 percent recorded for high-income countries. The automation gap is wide; the productivity gap is narrow.
| Share of jobs | Low- and middle-income countries, percent | High-income countries, percent |
| Exposed to automation risk | 4.5 | 14.2 |
| Could see meaningful productivity gains | 16.2 | 18.7 |
The two categories are distinct and neither is a jobs forecast. In the bank’s framework a task exposed to artificial intelligence can result either in automation, meaning the technology replaces a human in that task, or in augmentation, meaning it complements one. The first pair of figures measures the automation share and the second the augmentation share. Exposure signals potential for change, not job loss.
The bank frames the findings against what it calls the weakest average growth performance in three decades for developing economies. The report itself does not attach a growth rate to that description, and none is asserted here.
Constraints are treated as the binding issue rather than technology access. The bank notes that in Sub-Saharan Africa nearly one-third of rural schools still lack reliable electricity and more than two-thirds lack dependable internet access. Its prescription runs in three steps: adopt existing tools, adapt them to local conditions, then advance toward frontier development. Alongside that it calls for investment in electricity, connectivity, skills and institutional capacity, wider access to computing infrastructure, better availability of local data in local languages, conditions that let firms obtain finance and scale what works, and the building of public trust.
Indermit Gill, Senior Vice President and Chief Economist of the World Bank Group, said artificial intelligence has given developing economies a valuable opening, and that smaller and inexpensive tools adapted to local needs in health, education, legal services and agricultural advice can be put to work now, but that speed matters. Gaurav Nayyar, Director of the World Development Report 2026, said the window to get this right is narrow.
Why it matters: the report contains no Gulf Cooperation Council figure, and none is implied here. What it does supply is a framework the Gulf can read against its own position. Gulf states sit within the high-income group, where the electricity, connectivity and computing constraints the bank identifies as binding on developing economies are substantially less so. On our reading, that places the region on the adopt-and-adapt path from an infrastructure base most of the report’s subject economies do not have, and shifts the practical question from access to absorption, meaning skills, data in Arabic and institutional readiness. Kuwait’s diversification and human-capital agenda maps directly onto that second question.
Looking ahead: the report’s own framing is that the opening is time-limited rather than permanent. The measurable tests over the coming year are compute access, local-language data availability and firm-level adoption, all three of which the bank identifies as the constraints that decide whether the 16.2 percent productivity share is realised.
Sources: World Bank Group.
