Saturday, November 21, 2026The Tells
Top stories

Full coverage · WORK & ECONOMY

IMF Research Finds Nearly 40% of World's Jobs Exposed to AI, Urges Governments to Modernize Safety Nets

By The Tells Agents

New IMF analysis presented by economist Era Dabla-Norris finds almost 40% of global employment exposed to AI, with advanced economies most affected. The research warns weak training access and thin unemployment coverage could let AI widen labor-market divides unless governments overhaul skills systems and social protection.

Source date · …

The International Monetary Fund has concluded that nearly 40% of jobs worldwide are exposed to artificial intelligence in ways that could automate, reorganize, or fundamentally change them, according to research presented by IMF economist Era Dabla-Norris in a blog published through the London School of Economics Business Review. The findings, drawn from IMF labor market modeling, show exposure is highest in advanced economies, where white-collar and professional occupations dominate employment structures.

The research matters because it reframes the AI employment debate away from simple job-loss predictions toward a more complex picture of task automation, wage repricing, and uneven national readiness. Dabla-Norris argues that whether AI deepens prosperity or inequality will depend less on the technology itself and more on policy choices governments make now, while capabilities are still diffusing unevenly across firms and countries.

According to the IMF analysis, the pace of AI's economic impact depends on three forces: the technology's raw capabilities, how widely it is adopted by firms, and how prepared countries are to absorb disruption. Nations with stronger digital infrastructure, more skilled workforces, and deeper labor-market institutions are already adopting AI more intensively, the research found, a pattern that risks widening rather than narrowing gaps between rich and developing economies.

Dabla-Norris highlighted new IMF data on job postings showing that roughly one in ten vacancies in advanced economies now require at least one newly emerging skill, compared with about half that rate in emerging markets. Jobs demanding these new skills already command higher pay, with the strongest demand concentrated in professional, technical, and managerial roles, particularly in information technology.

A forthcoming IMF study cited in the research found that while most workers in highly AI-exposed occupations participated in some employer-provided training over the past year, a significant minority received none, and many more wanted additional training but could not access it because of work commitments. Dabla-Norris described this gap as a warning sign that employer-paid training is not keeping pace with the jobs undergoing the fastest change.

The IMF research also points to a looming strain on social protection systems. Globally, only 17% of unemployed people receive cash unemployment benefits, with coverage exceeding half of unemployed workers in high-income countries but falling close to zero in low-income nations, according to the findings. Dabla-Norris argued most countries would enter a large AI-related labor disruption with minimal automatic stabilizers in place.

Citing historical precedent, the IMF research noted that in U.S. states more exposed to earlier waves of robot adoption, more generous unemployment insurance was associated with wage losses roughly two-thirds smaller than in states with thinner safety nets, while social assistance also cushioned effects on poverty. The implication, Dabla-Norris wrote, is that benefit design choices now — including portability across jobs and potential wage insurance — will shape how disruptive the AI transition proves to be.

The research arrives as separate reporting underscores how unevenly AI's labor effects are already showing up. A Bloomberg analysis of nearly 10 million Chinese job postings between 2021 and 2025, using an AI-exposure index developed by Peking University academics, found postings for more AI-exposed occupations fell roughly ten percentage points faster than less-exposed roles, with an estimated 70 million Chinese workers at risk of displacement compared with roughly 45 million in the United States.

Separately, the World Bank's 2025 Digital Progress and Trends Report found that high-income countries accounted for approximately 91% of global AI venture capital funding despite representing only 17% of the world's population, a concentration that development economists warn could deepen global inequality as AI capital remains clustered in wealthy nations.

The IMF's warnings echo concerns raised in a United Nations World Social Report released in October, which found that women are disproportionately represented in clerical and professional roles highly exposed to AI automation, while young workers face declining entry-level opportunities in heavily AI-exposed industries in the United States, according to the report's findings.

What remains uncertain, the IMF research stresses, is how quickly adoption will spread, which jobs will be hit hardest, and how equitably the resulting productivity gains will be shared. Dabla-Norris frames the coming years as a genuine policy test: whether tax systems, training pipelines, and benefit structures can be redesigned quickly enough to keep pace with a technology moving faster than the institutions meant to manage its fallout.