A study published by MIT researchers in May 2026 examines a pattern that has held for centuries: technology creates jobs, but those jobs tend to go to young, skilled workers. The question the study poses is whether AI will follow the same pattern or break it — and the answer may determine the economic trajectory of the next decade.

Key Takeaways

  • The MIT study traces the historical pattern where new technologies create jobs that favour younger, more recently educated workers.
  • AI differs from previous technologies in three key ways: it automates cognitive tasks, it is being deployed faster, and it may erode the very skills workers need to adapt.
  • The optimistic view holds that AI will create new categories of work, as previous technologies did.
  • The cautious view is that AI may concentrate the benefits of automation more narrowly than previous technologies.
  • The outcome depends more on social and institutional choices — education, training, economic redistribution — than on the technology itself.

The Historical Pattern

The study, led by researchers at the MIT Department of Economics and the MIT Initiative on the Digital Economy, traces the relationship between technological change and employment across multiple industrial revolutions. The pattern is consistent: new technologies eliminate some jobs, create others, and reshape the skills that employers value.

In the Industrial Revolution, factory automation displaced skilled artisans but created new roles for machine operators, engineers, and factory managers. In the computer revolution, clerical and administrative jobs were automated, but new roles in software development, data analysis, and information technology emerged. In each case, the workers who benefited most were those who could adapt quickly — typically younger workers with more recent education and training.

The study's contribution is to quantify this pattern. Using historical data spanning two centuries, the researchers show that the "skill-biased technological change" observed in recent decades is not a new phenomenon. It is a pattern that has repeated across multiple technological transitions.

What Makes AI Different

The study identifies three factors that may make the current transition different from previous ones:

**Cognitive automation**: Previous technologies primarily automated physical or repetitive tasks. The steam engine automated muscle power. The computer automated calculation and record-keeping. AI, by contrast, automates cognitive tasks — writing, analysis, translation, programming, and even aspects of creative work. This affects a broader range of white-collar workers, including professionals who were previously insulated from automation.

**Speed of deployment**: AI tools are being adopted faster than previous technologies. ChatGPT reached 100 million users in two months, and enterprise AI adoption is accelerating. Workers have less time to adapt, and the education and training systems that typically help workers transition may not be able to keep pace.

**Skill erosion**: Some AI tools do not just automate tasks; they replace the need to develop certain skills. If AI writes code, junior developers may never learn to write it well. If AI translates, language learners may not persist. If AI summarizes documents, readers may not develop the ability to evaluate sources. This creates a "skill trap": the tools that help workers be productive in the short term may prevent them from developing the expertise they need to adapt in the long term.

The Broader Evidence Base

The MIT study is part of a larger body of research on AI and employment. The OECD has published several reports on the potential impact of AI on jobs, finding that while AI is unlikely to lead to mass unemployment, it will significantly reshape the distribution of work. The World Economic Forum's Future of Jobs Report estimates that AI will displace 85 million jobs by 2025 but create 97 million new ones — a net positive, but one that requires significant reskilling.

The International Monetary Fund has warned that AI could increase inequality, particularly in advanced economies, where workers in high-skill occupations may benefit while those in routine cognitive roles face displacement. The IMF's analysis suggests that AI could widen the gap between workers who can use AI effectively and those who cannot.

What This Means for Readers

Whether you are entering the workforce, changing careers, or helping young people think about their futures, the key takeaway is that the skills that matter most may be the ones AI cannot easily replicate: critical thinking, contextual judgment, empathy, and the ability to evaluate AI-generated output. The workers who thrive in an AI-influenced economy will be those who can use AI as a tool rather than being replaced by it.

The MIT study suggests that the critical factor is not whether AI will create jobs, but whether the education and training system can adapt quickly enough to prepare workers for the jobs that emerge. This is not a new question, but the speed of AI deployment makes it more urgent.

Age of Algorithms Perspective

The question of whether AI will follow the same historical pattern as previous technologies is not just an economic question — it is a question about power and distribution. Previous technological transitions created winners and losers, and the winners were typically those who already had access to education, capital, and social networks. AI may concentrate the benefits of automation more narrowly than previous technologies, not because of any inherent characteristic of the technology, but because the institutional mechanisms for distributing the benefits — strong unions, progressive taxation, robust public education — are weaker than they were during previous transitions.

The outcome of the AI transition will depend less on the technology itself and more on the social and institutional choices we make. That is both a warning and an opportunity. The technology is not deterministic. The choices are ours.