Artificial intelligence (AI) could augment, downgrade or eliminate about 40% of jobs globally, International Monetary Fund Managing Director Kristalina Georgieva said, warning that young people entering the workforce face particular risks as automation changes hiring patterns.
Speaking at the 2026 World Bank Group and IMF Annual Meetings in Bangkok, Georgieva said the effects of AI are already emerging in labour markets, with demand shifting away from routine, repetitive work towards jobs requiring more advanced skills.
The IMF estimates that about 60% of jobs in advanced economies are exposed to AI, compared with 23% in low-income countries. Georgieva cautioned that the lower exposure in poorer economies does not necessarily indicate greater protection, saying it partly reflects the fact that AI adoption has not yet reached those countries at the same pace.
“What it is doing is it is shifting demand for people now,” Georgieva said, adding that the longer-term effects remain uncertain.
The changing pattern presents a challenge for young workers, who traditionally entered employment through routine jobs before progressing into more skilled and better-paid positions. If those entry-level opportunities shrink, new graduates could struggle to gain the experience needed to advance their careers.
Georgieva cited an IMF study of the United States showing that jobs requiring AI skills already account for 10% of employment opportunities in the area examined. Those positions tend to pay better, while routine and repetitive jobs are declining.
The same study found growth in lower-paying employment, including construction linked to data centres and service jobs. Although the net employment effect in the area studied was positive, Georgieva said the gains were concentrated in lower-paid work, complicating the outlook for living standards.
The economic implications extend beyond hiring. As AI changes the balance between labour and capital, governments may face pressure to reconsider how they raise revenue to finance public services and help workers adapt.
“Why? Because traditionally, governments would generate revenues by taxing, mostly, people,” Georgieva said, questioning how governments would maintain revenues if labour’s share of economic activity declines relative to capital.
She called for better labour-market data, more responsive education systems and lifelong learning to help workers update their skills as technology evolves.
Singapore President Tharman Shanmugaratnam said the decline in demand for first jobs was already becoming apparent, particularly in technology-related sectors. He urged governments to work with businesses to create pathways into employment for graduates.
Tharman also warned that AI could deepen a K-shaped labour market in which highly skilled workers benefit while the broad middle of the workforce loses opportunities.
The policy challenge, the speakers argued, is to ensure that productivity gains from AI translate into better employment and income growth rather than a widening divide between workers who can use the technology and those who cannot.
