AI and other new digital skills are appearing in roughly one in ten vacancies in advanced economies and pay a clear wage premium. Yet their spread is linked to sharper labor-market polarization—helping high-skilled workers while hollowing out middle-skilled roles and reducing employment in AI‑exposed occupations with low worker complementarity, with young workers hit especially hard.

Main Finding

New (especially IT and AI) skills are rapidly reshaping labor markets. Roughly 1 in 10 job vacancies in advanced economies—and about half that rate in emerging economies—now request at least one new skill. These skills raise wages and employment on average but increase labor-market polarization, disproportionately benefiting high-skilled workers and (indirectly, via higher service demand) some low-skilled workers while eroding middle-skilled employment. Vacancies requiring AI skills pay premia, yet AI diffusion is associated with lower employment in occupations that are both highly exposed to AI and have low complementarity with it, with pronounced effects for young workers. Cross-country variation in demand versus supply of new skills is large, captured by a Skill Imbalance Index, implying different policy priorities across economies.

Key Points

• Incidence
  • ~10% of vacancies in advanced economies request at least one new skill; about 5% in emerging economies.
  • New-skill requirements tend to emerge first and most intensely in the United States.
• Wage and employment effects
  • Vacancies demanding new skills (including AI) offer higher wages on average.
  • Overall employment and wages rise where new skills are adopted, but gains are uneven.
• Polarization and distributional consequences
  • Labor-market polarization intensifies: gains concentrated among high-skilled workers.
  • Low-skilled workers can benefit indirectly through increased demand for services supplied to high-skilled earners.
  • Middle-skilled occupations are most at risk, contributing to a shrinking middle class.
• AI-specific patterns
  • AI-skill vacancies carry wage premia.
  • Diffusion of AI skills corresponds to lower employment in occupations with high exposure to AI and low complementarity (i.e., tasks that AI can perform without close worker complementarities).
  • Young workers face particular challenges in occupations exposed to AI.
• Cross-country heterogeneity
  • A Skill Imbalance Index—measuring relative demand versus supply of new skills—shows wide variation across countries.
  • Countries with strong demand for new skills should emphasize education and reskilling; countries with strong supply should focus on firm absorption (innovation, financing, technology adoption).

Data & Methods

• Data sources (as described)
  • Vacancy-level data across a set of advanced and emerging economies.
  • Identification of skills through text analysis of job postings (skills labeled as "new" include IT and AI-related competencies).
• Measurement
  • Incidence: fraction of vacancies requesting at least one new skill.
  • AI-skill flag: vacancies explicitly requesting AI-related competencies.
  • Skill Imbalance Index: constructed to capture the gap between skill demand (vacancies) and skill supply (workers/skill endowments or related proxies) at the country level.
  • Occupation-level exposure and complementarity: occupations classified by their technical exposure to AI and the degree to which worker tasks are complementary to AI.
• Empirical strategy (high level)
  • Vacancy-level regressions estimate wage premia associated with new-skill requirements, controlling for occupation, firm, and other observables.
  • Cross-sectional and panel analyses relate the diffusion of new skills to changes in employment across occupations and demographic groups (notably youth).
  • Country-level comparisons use the Skill Imbalance Index to link macro patterns to policy prescriptions.
• Robustness and limitations (implicit)
  • Results rely on accurate skill extraction from vacancy texts and on valid measures of occupational exposure/complementarity.
  • Causal interpretation of diffusion effects may be limited by endogeneity (e.g., technology adoption responding to labor-market conditions).

Implications for AI Economics

• Policy targeting depends on national context
  • For economies with strong demand for new skills: scale up education, lifelong learning, and targeted reskilling programs to meet employer needs and avoid bottlenecks.
  • For economies with strong supply but weak demand: focus on policies that enable firms to absorb skills—encourage innovation, reduce barriers to firm investment in technology, and improve access to credit.
• Labor-market and social policy
  • Anticipate and mitigate polarization: reinforce apprenticeships, mid-career training, and transition support for displaced middle-skilled workers.
  • Protect vulnerable groups (especially youth) with targeted entry-level training and pathways to higher-complementarity tasks.
  • Consider redistribution and social insurance instruments to manage transitional inequality as skills reallocate returns across workers.
• Firms and industry
  • Firms should invest in organizational complementarities (task redesign, team composition) to capture productivity gains from new skills without displacing workers unnecessarily.
  • Small and medium enterprises may need public support to adopt AI/IT and to utilize local skill supplies.
• Research directions for AI economics
  • Better measurement of skill supply (worker upskilling trajectories) and causal evaluation of training programs.
  • Deeper analysis of complementarities between AI and specific occupations/tasks to predict employment dynamics more precisely.
  • Cross-country causal studies on how policy levers (education, innovation finance, labor-market regulation) shape the demand–supply balance for new skills.