The managing director of the International Monetary Fund, Kristalina Georgieva, used stark language in Davos on Friday, warning that artificial intelligence is sweeping through labour markets like a “tsunami” and that younger workers will be hit hardest. IMF research presented at the World Economic Forum projects that in the coming years some 60% of jobs in advanced economies and roughly 40% worldwide will be affected by AI — through enhancement, replacement or reshaping of tasks.
Georgieva noted that about one in ten jobs in advanced economies has already been “augmented” by AI, generating higher pay and positive spillovers for local economies. Yet she stressed that many of the roles being eliminated are traditional entry-level positions that young people rely on to gain workplace experience, leaving new labour-market entrants with far fewer stepping-stone opportunities.
The warning came amid broader debate in Davos about how the gains from AI will be distributed. Christy Hoffman, secretary-general of the global services workers’ union, argued that firms deploy AI principally to raise productivity and cut costs — and that will often mean layoffs unless policy buffers are put in place. Microsoft’s CEO Satya Nadella cautioned that if AI fails to deliver social value beyond a handful of tech giants, public tolerance for its resource use and market power will erode.
Georgieva’s principal anxiety was the speed of technological change outpacing governance: regulators have not yet developed robust frameworks to ensure AI’s safety or its inclusiveness. European Central Bank president Christine Lagarde added a geopolitical wrinkle, warning that rising protectionism and fractured cooperation over data and capital could choke off the very investments and inputs (energy, data, finance) an AI-driven economy needs.
On the technical front, Demis Hassabis of Google DeepMind rejected claims that the industry has already passed into a post-AGI era, saying substantial breakthroughs — such as persistent continual learning and far stronger memory systems — are still required and that a five- to ten-year horizon to true artificial general intelligence remains plausible. The technical timeline matters because the nearer-term disruption does not require AGI: current narrow AI systems are already reshaping many occupations.
The debate in Davos underscores a wider policy choice. Governments and firms face a trade-off between rapid adoption of productivity-enhancing tools and the social costs of dislocated workers, particularly inexperienced entrants to the labour market. The IMF’s figures make it clear that without active labour-market policies, targeted retraining, and redistribution of productivity gains, the next decade could see rising inequality, stalled social mobility and political backlash against both firms and regulators that permit unfettered automation.