OPEN
AI will affect almost 40 percent of jobs around the world, replacing some and complementing others.
Source: International Monetary Fund ·
Analyst note
Georgieva’s statement is easy to misread if “impact” becomes “replace” in the retelling. The IMF’s own follow-ups stress distributional effects and policy design; vendors stress inevitability; unions stress deskilling and monitoring risks.
By 2030, the key empirical disagreement is whether automation shows up as (a) unemployment, (b) wage compression in exposed occupations, or (c) hidden productivity gains inside incumbent firms. Model releases from major labs affect beliefs faster than they affect measured stocks.
Evidence timeline
The IMF framed global labor exposure to AI as substantial, while emphasizing heterogeneity across countries and income levels.
Labor economists debated interpretation: ‘impact’ is not identical to displacement, and complementarity channels matter for net employment.
Task-level reallocation evidence accumulated through 2025–2026, but clean attribution of net employment shifts to AI (rather than rates, fiscal policy, or sector cycles) remained methodologically hard.