From exposure to opportunity: Why skills shape the employment effects of new technologies

Much of the debate about new technologies and the future of work starts with potential exposure. Which occupations can AI transform? Which jobs are most at risk?

These are important questions. Recent ILO estimates suggest that one in four workers worldwide is in an occupation with some exposure to generative AI. But exposure is not the same as job loss. Because most occupations still contain tasks requiring human input, transformation rather than replacement is considered the more likely outcome (Gmyrek et al., 2025).

This raises another question. If exposure tells us where technology could affect jobs, what determines whether that change translates into more or fewer employment opportunities?

Our research at the ILO indicates that skills are part of the answer.

The same exposure, different outcomes

Using ILO Harmonized Microdata for more than 1,000 subnational areas across 69 countries, we examine how employment changes in areas with different levels of potential exposure to 40 emerging digital technologies (Delaporte et al., forthcoming). Our analysis compares areas at the first administrative level within countries – such as states, provinces or regions – and measures changes in their employment-to-population ratio.

Exposure captures how strongly the sectors in which people were initially employed are potentially affected by these technologies. It therefore measures potential exposure rather than actual technology adoption. These technologies range from industrial automation and robotics, e-commerce and digital payments to cloud computing, machine learning, and digital health technologies (Prytkova et al., 2024).

On average, we find that greater exposure leads to employment gains. But those gains are uneven. The share of people who are employed rises more among prime-age workers than among younger and older workers. The effects of technological exposure also differ substantially across countries’ income levels: in poorer economies, increases in employment are concentrated among lower-educated workers, while in richer economies they increasingly extend to highly educated workers.

These differences suggest that exposure alone cannot explain how employment responds to technological change. As previous research has stressed, its employment consequences depend on how technology is introduced and how workers, firms and institutions respond (Berg, 2024; Gmyrek et al., 2024). Our results point to another important factor that can shape that response: the skill mix embedded in local labour demand.

The skills embedded in local labour demand matter

We examine 15 different categories of skills – spanning cognitive, socio-emotional and manual skills – and measure how important each skill is in the mix of occupations in each area.

Some skills stand out. In areas with stronger demand for sophisticated cognitive skills, basic computer skills, financial skills, writing, and project and process management, exposure to emerging digital technologies leads to larger increases in employment. The same is true in areas where people management skills are more strongly demanded.

Importantly, these relationships remain even after accounting for differences in educational attainment.

This distinction matters. Education provides an important foundation for acquiring skills, but qualifications alone cannot capture the full range of skill sets workers actually use at work. Two areas can therefore have similarly educated workforces and still have very different occupational skill profiles – and these differences can shape how employment responds to technological change.

This is not an entirely new idea. Research has shown that technological change can increase the value of skills that complement rather than replicate what technology can do. Socio-emotional skills, for example, have become increasingly valuable as workplaces place greater importance on coordination and interaction between workers (Deming, 2017).

The ILO's Lifelong Learning and Skills for the Future report emphasizes that digital transformation requires broader and complementary skill sets alongside technical skills (ILO, 2026). Our results indicate that these broader skills also matter for how technological change translates into employment.

Not just technical skills

There is another finding that is worth paying attention to.

The skills associated with increases in employment are not necessarily the most technologically advanced ones.

Higher demand for machine learning and AI skills is, on average, associated with weaker employment responses to digital exposure. Demand for software and technical skills show a more mixed relationship, with their role varying considerably across countries and groups of workers.

This does not mean that these skills are unimportant. Economies clearly need people who can develop, deploy and maintain new technologies. But there is a difference between having the skills needed to create and operate an emerging technology and having the skills that help a much broader group of workers adapt when that technology changes their jobs (Delaporte and Liepmann, 2026).

For the latter, a wider set of skills may matter. Being able to solve new problems, use digital tools, communicate information, coordinate with others, manage projects and adapt may become increasingly important as tasks and working methods change.

This also echoes our earlier finding that emerging digital technologies are creating demand not only for technical expertise, but for a much broader set of skills (ILO, 2026). As technology changes jobs and tasks, it can also increase the importance of the other skills – cognitive, social and managerial – that workers use alongside new technologies.

From exposure to opportunity

These findings challenge the idea that there is a universal set of “future-proof” skills or that highly sophisticated technical skills alone are the solution. The skills associated with increases in employment vary across countries’ income levels and across groups of workers.

This has important implications for how we prepare workers for technological change.

Preparing workers cannot simply mean identifying the latest technology and training everyone in the technical skills associated with it. Nor is increasing educational attainment, important as it is, enough on its own.

Skills strategies need to pay attention to the broader mix of skills that complement new technologies. That means combining technical and digital expertise with cognitive, communication and managerial skills that can help workers adjust as their tasks change. And because those tasks will continue to evolve, workers need opportunities to acquire and update these skills throughout their working lives (ILO, 2026).

But skills and lifelong learning are only part of the response. Technological change will inevitably create difficult transitions for some workers, including displacement and job loss. Social protection therefore remains essential to provide income security during these transitions and give workers the support and time they need to move towards new employment opportunities.

Measures of technological exposure have greatly improved our understanding of where change may occur. Our findings suggest that the skill mix embedded in local labour demand helps shape how workers are positioned to respond. The challenge now is to translate that knowledge into policies that combine opportunities to develop and update skills with the protection workers need to navigate technological change – and ultimately help turn technological exposure into opportunity.