Why Workers in Rich Countries Face More AI Risk Than Those in Developing Economies

AI will not disrupt jobs evenly across the world: workers in high-income countries are far more exposed because they perform more digital, cognitive tasks within the same occupations. The biggest drivers of AI exposure are technology use and skills, not job titles, meaning global inequality will shape who feels AI’s impact first.

As artificial intelligence rapidly enters offices, factories, and service industries, a common fear dominates public debate: that AI will disrupt jobs everywhere, all at once. But new research from the Institute for Structural Research (IBS) in Poland and the Asian Development Bank (ADB) paints a very different picture. The global workforce, the study shows, is far from equally exposed to AI. In fact, where a worker lives matters just as much as what job they hold.

Using data from more than 100 countries, the researchers find that workers in low-income countries are far less exposed to AI than those in richer economies. The difference is large, nearly a full standard deviation on average. This gap helps explain why AI anxiety feels more intense in some countries than others, and why fears of a universal global jobs shock may be overstated.

It’s About Tasks, Not Job Titles

Most studies measure AI exposure by looking at occupations, such as clerks, accountants, and managers, and assuming those jobs look similar everywhere. This paper challenges that assumption. Instead of focusing on job titles, the authors examine what workers actually do each day.

Using detailed surveys from the OECD, the World Bank, and China, the researchers analyze tasks such as reading and writing, problem-solving, managing information, using computers, and making independent decisions. These are the activities modern AI systems are increasingly capable of performing.

The results are striking. Workers with the same job title often perform very different tasks depending on their country. A manager or technician in a high-income economy typically spends much more time on abstract thinking, communication, and data processing than someone in the same occupation in a poorer country. Those differences, more than job structure itself, determine exposure to AI.

Why Rich Countries Face Higher AI Exposure

The study finds that nearly 80% of global differences in AI exposure are driven by task differences within occupations, not by the mix of jobs in each country. High-income countries consistently show higher exposure because workers there perform more AI-relevant tasks.

The single most important factor is access to digital technology. In countries and sectors where computers are widely used, workers are far more exposed to AI. Once digital tools become part of everyday work, exposure rises sharply.

Education and skills also matter. Workers with higher levels of education and stronger literacy skills face greater AI exposure because their jobs involve more reading, writing, analysis, and coordination. Even within the same occupation, better-skilled workers are more likely to perform tasks that overlap with AI capabilities.

Globalization has a more mixed effect. In many developing countries, participation in global value chains, especially in agriculture, mining, or basic manufacturing, is linked to lower AI exposure, because these sectors rely on routine tasks that AI currently struggles to perform.

AI Exposure Is Rising, Quietly

The researchers also examine how AI exposure has changed over time in advanced economies. Using two waves of international survey data, they find that exposure has increased steadily since the early 2010s.

Importantly, this rise is not driven by people changing jobs. Instead, the nature of work within jobs is shifting. Office workers, professionals, and technicians are increasingly doing tasks, like information processing and digital communication, that align closely with AI capabilities, even when their job titles remain the same.

This quiet transformation suggests that AI’s impact on work is gradual and subtle, reshaping tasks long before it reshapes employment numbers.

A World Divided by AI Readiness

When AI exposure is combined with employment data, a clear global divide emerges. High-income countries, which employ just over a quarter of the world’s workforce, are home to nearly 60% of the most AI-exposed workers. Low- and lower-middle-income countries, by contrast, contain the majority of the least exposed workers.

The authors stress that exposure does not mean job loss is inevitable. Whether AI replaces or supports workers depends on how firms adopt the technology, how expensive automation becomes, and how work is organized. Still, exposure shows where AI is most likely to matter first.

The message is clear and sobering. Artificial intelligence is not sweeping across the global labor market in a single wave. It is advancing unevenly, shaped by technology, skills, and development. Those differences will determine who feels AI’s benefits and its disruptions first.