ILO and World Bank paper says GenAI may deepen labour-market divides
New research from the ILO and World Bank finds that GenAI disruption could outpace productivity gains in poorer economies.
A joint working paper by the International Labour Organization (ILO) and the World Bank says generative AI is likely to reshape labour markets globally, but not in the same way across countries.
The paper finds that advanced economies face greater overall exposure, while developing economies may see disruption arrive faster than productivity gains due to weaker digital infrastructure and differences in how work is organised.
Prepared as a background study for the World Development Report 2026, the paper examines labour-market exposure to GenAI across 135 countries, covering about two-thirds of global employment. According to the study, digital infrastructure and job-task composition are among the main factors shaping the distribution of risks and opportunities between advanced and developing economies.
Exposure is highest in advanced economies, especially in clerical and professional occupations. Lower-income countries are less exposed overall, but the paper says structural constraints reduce their ability to benefit from the technology. A central concern is that workers in jobs vulnerable to automation are often already online, even in poorer settings, meaning displacement could happen relatively quickly.
The paper also says many of the jobs most exposed to automation in developing economies are relatively higher-quality roles, including clerical and administrative work that has often provided a route into decent employment, especially for women and young workers. AI-driven automation, the study warns, could narrow those pathways.
Potential gains are also uneven. Many workers in jobs that could benefit from GenAI lack reliable internet access in lower-income settings. The paper adds that the same occupation title can involve different tasks depending on the country, with workers in poorer economies often carrying out fewer non-routine analytical tasks, relying less on computers, and doing more routine or manual work. Such differences reduce the scope for productivity gains from GenAI deployment.
ILO and the World Bank conclude in the paper that GenAI’s labour-market effects will depend not only on the technology itself, but also on digital connectivity, skills, task organisation, labour-market institutions, and social protection. Expanded digital access, stronger skills policies, and better labour protections are presented as necessary if the gains from GenAI are to be shared more broadly.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!