This ILR Blog explores how AI and automation are reshaping self-employment across Europe. It shows that while technology can create new opportunities for some, it also risks pushing others – especially low-paid workers – into precarious solo work.

Whether you’re a lawyer working in a law firm and thinking about starting your own practice, a cleaner navigating unpredictable hours and looking for a more stable job, or a programmer weighing the freedom of freelancing against the stability of a tech job — your choice between being self-employed or working for someone else is shaped by many factors. Research shows that personal preferences and broader economic and social conditions play an important role. What’s less obvious is how the technology that you use in your job — like AI tools or task automation — might be tipping the scales.

This is what we explore in our study.(1) We wanted to understand how technology — especially AI and automation — affects people’s real-life decisions to move between dependent employment and self-employment, and whether certain worker groups (differing in skills, incomes, or age) are affected differently. We tracked workers’ transitions been employment, solo self-employment, and self-employment with employees across 31 European countries.(2) This allowed us to see how exposure to different types of technology affects individuals’ career paths — whether people stay in employment, become self-employed, or move back and forth as technological progress unfolds.

Defining self-employment, we distinguish between solo self-employment (working independently, without employees) and self-employment with employees (running a business with staff). With respect to technology, we looked at two types: AI — technology that often augments work, helping people do more and better work; and automation — technology that tends to replace tasks, especially repetitive ones.

We found that people exposed to AI in their jobs are more likely to enter solo self-employment, but they are also more likely to return to traditional employment. This suggests that AI can create new possibilities for independent work, especially for those who have the skills and flexibility to take advantage of it. But it could also imply that solo self-employment may not be desirable, especially in the long term. Our analysis shows that such risks are particularly pronounced for low-skilled workers and workers with low wages.

By contrast, workers in occupations where automation technologies play an important role are less likely to become self-employed. It seems that when automation limits the kinds of tasks humans are needed for, it also reduces the chances of building a viable solo business in those sectors.

This suggests that technology may open new doors for highly skilled workers — but for others, it can mean losing job security and having to go it alone, often with limited support. Many of these workers do not become self-employed by choice, but because they cannot find decent, stable employment.

This challenges the idea that technology always empowers. Rather, there is a danger that technology increases inequality in who gets to benefit. Without smart policies, we risk a world where digital transformation leaves too many people behind.

Our findings point to several areas where policy action is needed:

1. Skills and digital inclusion: Workers need support to adapt to new technologies. That means targeted training — especially for those in low-skilled or vulnerable jobs — to help them stay employable or start sustainable businesses.

2. Modern social protections: Many self-employed workers fall outside traditional safety nets. Social protection systems need to reflect the reality of today’s workforce, including non-standard and solo self-employment.

3. Support for decent work: It is not enough to celebrate flexibility or entrepreneurship. We need to ensure that self-employment — especially solo self-employment — can be a pathway to decent, secure work, not just a fallback when no other jobs are available.

Self-employment is often seen as a sign of innovation and flexibility. And sometimes it is. But that’s not the whole story. Workers with high skills who are exposed to AI can choose whether to stay employed or become entrepreneurs. Others — often in low-paid or routine jobs — may be pushed into solo self-employment because of the lack of good alternatives in the regular labour market.

Taking action to ensure that policies help workers’ career transitions matters not only for individual workers, but for social cohesion and the long-term sustainability of labour markets and welfare systems.

Future research should analyse whether the trends we uncover with respect to AI continue or even intensify as AI adoption spreads across the economy and the labour market. In this context, analyses that examine the effects of differences in the regulatory framework between countries seem particularly important.

(1) Bachmann, R., M. Gonschor, A. Mitra, S. Milasi (2025), Technological progress and the dynamics of self-employment: Worker-level evidence for Europe. International Labour Review 164(2): 1-23.

(2) Our analyses are based on data from the European Union Statistics on Income and Living Conditions (EU-SILC) for the time period 2014-19.