AI and Labor: Who’s Most at Risk in High, Middle, and Low-Income Countries?

The study by the World Bank's Human Capital Project analyzes AI's impact on labor markets in low- and middle-income countries, revealing stark disparities in AI exposure based on income levels, education, and infrastructure. While high-income nations lead AI adoption, developing economies face barriers like limited electricity and internet access, requiring targeted policies to bridge the digital divide and ensure inclusive AI-driven growth.

Research published by the World Bank's Human Capital Project and authored by Gabriel Demombynes, Jorg Langbein, and Michael Weber, examines the effects of AI on labor markets in developing economies. While previous research has focused mainly on high-income countries, this study expands the discussion by analyzing labor force microdata from 25 countries covering 3.5 billion people. It employs the AI Occupational Exposure (AIOE) index, developed by Felten et al. (2021), to assess how AI exposure varies across different income levels, industries, and worker demographics. The results highlight significant disparities, with high-income countries (HICs) leading in AI exposure, while low-income countries (LICs) remain the least affected due to differences in economic structures, workforce composition, and infrastructure.

Who is Most Exposed to AI?

A sharp contrast emerges when looking at AI exposure by income level. High-income countries have the greatest exposure to AI, with an average index score of 62, followed by upper-middle-income countries at 49, lower-middle-income countries at 44, and low-income countries at 37. This reflects the predominance of AI-intensive occupations in wealthier economies, where white-collar jobs are more common. The study also identifies clear socio-demographic trends in AI exposure. Women are more exposed to AI than men across all income levels, primarily because they occupy clerical and administrative roles, which are highly susceptible to automation. AI exposure rises with education levels, particularly in middle-income nations, where better-educated workers are more likely to be in AI-affected jobs. Urban workers face significantly higher AI exposure than rural workers, emphasizing a digital divide that widens economic disparities. Interestingly, younger workers (aged 15-24) have lower AI exposure, as they are more likely to be engaged in manual, informal, or entry-level jobs that AI is less likely to disrupt in the near term.

AI's Impact on Different Sectors and Jobs

The study provides valuable insights into which industries and occupations are most affected by AI. High-skilled occupations, such as managers, professionals, and technicians, have the highest AI exposure, while clerical jobs are at particularly high risk due to their reliance on structured, repetitive tasks that AI can automate. In contrast, blue-collar and manual labor jobs, such as those in agriculture, construction, and manufacturing, exhibit minimal AI exposure, reinforcing the notion that AI’s immediate impact will be concentrated in white-collar professions. Industry-wise, commerce, financial services, and public administration have the highest AI exposure, while construction, mining, and agriculture remain largely untouched. These findings suggest that AI will likely reshape service-oriented sectors before affecting industrial and manual labor markets.

The Infrastructure Divide: A Barrier to AI Adoption

One of the most striking findings of the study is that basic infrastructure limitations in low-income countries prevent AI from having an immediate impact on their labor markets. 41% of AI-exposed occupations in LICs lack reliable electricity, rendering AI integration almost impossible. The gap is even more pronounced in rural areas, where 51% of workers face AI exposure but have no access to electricity. Even in lower-middle-income countries, 5% of AI-exposed occupations still lack basic power infrastructure. Without consistent access to electricity and the internet, AI advancements will remain largely theoretical in these economies. This highlights a critical challenge: while high-income countries are already adopting AI-driven automation, many developing nations lack the infrastructure to make AI a viable tool for economic transformation.

Bridging the AI Gap: The Path Forward

While AI is often portrayed as a force that will disrupt labor markets, the study emphasizes that AI can also enhance productivity and improve service quality, particularly in healthcare and education. In low- and middle-income countries, AI can be leveraged to improve access to medical services, enhance learning outcomes, and streamline government operations. However, the study warns that without targeted policy interventions, AI could exacerbate global economic inequalities. The research suggests five key policy recommendations to ensure that AI benefits are distributed equitably:

• Invest in AI-related education and digital skills to equip workers with competencies for an AI-driven future.
• Expand electricity and internet infrastructure, especially in rural areas, to enable AI adoption where it can have the most impact.
• Focus on AI augmentation rather than automation, ensuring that AI complements human labor rather than replacing it.
• Encourage AI-driven innovations in healthcare and education, where AI can help bridge human capital shortages.
• Address the urban-rural divide by ensuring that AI adoption does not widen existing inequalities.

The study challenges the prevailing narrative of AI-induced job losses by demonstrating that low-income countries are unlikely to experience immediate labor market disruptions from AI due to structural economic factors. However, middle-income countries are already experiencing shifts, particularly in digitizing service sectors. Without proactive strategies, the digital divide between rich and poor nations could widen, making it even more difficult for lower-income countries to catch up with global economic trends. The study underscores that AI’s impact is not universal it depends on a country’s economic structure, workforce composition, and technological readiness.

Overall, this research provides a critical early benchmark for discussions on AI investments, labor market policies, and economic development strategies in low- and middle-income countries. Policymakers must ensure that AI is not just a tool for automation but a driver of inclusive economic growth. By recognizing AI’s diverse implications and taking proactive steps to integrate AI responsibly, developing economies can harness its potential while safeguarding workers' rights and livelihoods.