From Automation to Skills: How AI Is Creating Jobs While Squeezing the Middle Class

The IMF finds that AI and digital technologies are transforming labour markets mainly by creating new skills rather than eliminating jobs, raising wages and employment overall but deepening job polarisation. While non-AI skills support job growth, rapid AI adoption risks reducing employment in highly exposed occupations, especially for young and middle-skill workers, unless education and training systems adapt quickly.

Prepared by researchers at the International Monetary Fund, using large-scale labour market evidence from Lightcast, the Organisation for Economic Co-operation and Development (OECD), the International Labour Organization (ILO), and national statistical agencies, this report shows how artificial intelligence and digital technologies are reshaping jobs not through sudden mass layoffs, but through a fast and uneven shift in skills. Across advanced economies, about one in ten job vacancies now requires at least one “new skill” that barely existed a decade ago. In emerging markets, the figure is closer to one in twenty, reflecting differences in economic structure and access to technology.

These new skills include data analysis tools, cloud computing, cybersecurity, digital marketing platforms, and increasingly, AI-related applications. They are most common in professional, technical, and managerial jobs and are strongly linked to occupations that are expanding rather than shrinking. In simple terms, where new skills appear, jobs are more likely to grow.

Why Technology Is Creating Skills, Not Just Replacing Jobs

The report challenges the idea that technology mainly destroys work. Instead, it shows that technological change creates new tasks that require new skills. Information technology dominates this process, accounting for more than half of all emerging skills. Within IT, AI-related skills have grown rapidly and now make up nearly a third of new IT skills.

Importantly, most AI-related demand is not for elite programmers building advanced models. It is for workers who can use AI tools in everyday work, such as drafting content, analysing data, automating routine reports, or supporting decision-making. These “AI-user” skills now appear in many non-technical roles, from marketing and finance to health care and education, highlighting AI’s role as a general-purpose technology.

Higher Pay, Uneven Gains

New skills clearly pay off for many workers. Job vacancies that list new skills offer wages that are about 3 to 4 percent higher than similar jobs without them. When multiple new skills are required, the wage premium rises further. IT, engineering, and data-related skills deliver the strongest pay gains.

At the local economy level, regions with faster growth in new skills experience higher average wages and more employment overall. High-skilled workers benefit the most, but low-skilled workers also gain indirectly as rising incomes increase demand for services such as retail, care, and hospitality. However, middle-skilled workers, especially those in routine office and production jobs, see little benefit. This deepens job polarisation and contributes to the long-term shrinking of the middle class.

AI’s Special Challenge for Young Workers

While new non-AI skills boost employment, AI-related skills show a more mixed picture. Jobs that require AI skills tend to pay more, but they have not yet increased overall employment. In occupations that are highly exposed to AI and offer little opportunity for humans to complement machines, employment actually falls over time.

Young workers are particularly affected. Entry-level jobs often involve tasks that AI can automate easily, making it harder for young people to get a foothold in white-collar careers. Five years after AI skills enter a local labour market, employment in these high-exposure occupations is noticeably lower than in less exposed areas. The report warns that without targeted policies, AI could narrow career entry points even as productivity rises.

Who Is Ready, and What Governments Must Do

Countries differ significantly in their ability to meet the rising demand for skills. Advanced economies with strong education systems and high shares of professional jobs face the greatest demand for new skills. Some countries, such as Sweden and the Netherlands, struggle to supply enough trained workers. Others, such as Ireland and Poland, produce many skilled graduates but lack strong firm demand to absorb them.

To guide policy, the report introduces new indices measuring skill imbalances and readiness. The message is clear: there is no one-size-fits-all solution. Countries with skill shortages should expand training, integrate digital skills into all fields of study, and encourage labour mobility. Countries with strong skill supply should focus on innovation, access to finance, and business growth.

Across all economies, the report calls for lifelong learning systems, fair competition policies, support for young workers, and stronger social protection. The AI age, it concludes, can deliver broad prosperity, but only if societies invest as much in people and skills as they do in technology.