A debate that recurs cyclically in economics is whether technology destroys jobs and generates unemployment. In the first part of this century, advances in automation and digitization, accelerated by the arrival of artificial intelligence (AI), raise their importance and lead us to talk more about social transformation than about a mere technological change that influences employment as we know it.
“Digitalization is redefining what type of employment is stable and what type of employment remains precarious”
In Spain, the data indicate that we are not experiencing a massive substitution of workers by machines in the style of the more alarmist predictions. But that does not mean that the characteristics of employment are not changing. What the data show is a quieter process, but potentially more substantial. Digitalization is redefining what kind of employment is stable and what kind of employment remains precarious. And that difference could, in the future, be more important than the total amount of employment.
If there is one hallmark of the recent evolution of the Spanish economy, it is the improvement in the level of employment after the pandemic. Unemployment rates have fallen and the labor market has shown unexpected resilience. However, as has already been noted in other analyses, reducing unemployment does not automatically imply greater long-term solidity.
The key lies in how that employment is being generated and whether it provides us with the possibility of maintaining the improvements. Here is where the role of digitalization enters. Not as a force that directly substitutes workers, but as a factor that reconfigures the way the economy and employment are organized. The impact shakes pillars of our society that go beyond the realm of economics and promotes contemporary phenomena unique such as the rise of absenteeism, telework, labor nomadism, or the gig economy (precarious for some) of brands like Glovo, Uber or Fiverr.
With the possibilities offered by AI-driven productivity gains, it is foreseen that in a couple of decades (if not sooner) we may be debating the voluntariness of paid work. We would thus be able to question the historical belief in work as mankind’s predestination to earn a living with the sweat of one’s brow.
The impact of digitalization in Spain: temporality and the weekly working hours
One of the clearest results is that digitalization does not significantly explain the evolution of employment or unemployment. After accounting for the economic cycle, measured by GDP growth, the impact of digital intensity on total employment is not decisive: there is no evidence that, in aggregate, we are witnessing a reduction in employment in Spain. This places the Spanish case closer to European evidence than to the results observed in the United States with industrial robotization in urban areas highlighted by Nobel laureate Daron Acemoglu.
“Sectors with higher digital intensity show a lower share of temporary employment, and the effect is especially pronounced in the Spanish case”
Where changes are indeed noticed is in employment stability and the type of contract. Sectors with higher digital intensity show a lower share of temporary employment, and the effect is especially strong in the Spanish case when compared with Europe.
This matters for two reasons. First, because Spain is one of the countries with the greatest labor duality in Europe, with a clear difference between the conditions of permanent and temporary workers. Second, because it suggests that digitalization is not associated with greater precariousness, but rather with the opposite: in the post-COVID years, the most digital sectors tend to generate more stable employment.
The digitalization also affects the working week. Once sectoral differences are controlled for, the more digitalized sectors tend to record slightly more hours worked per week.
This result also challenges the idea that technology reduces working time. In reality, what seems to have occurred in the last decade is a shift in the composition of employment, with fewer temporary jobs, more stable jobs and a higher work intensity in some sectors.
A dual labor market even in the digital realm: implications for economic policy
However, these effects are not homogeneous. Spain presents a clearly dual structure, with digital sectors (information and communications, professional services) offering greater stability and less temporariness, coexisting with less digitalized sectors (hospitality, construction) with higher temporariness and longer working hours.
This implies that digitalization is not transforming the entire labor market equally, but may be reinforcing an existing division. If digitalization generates more stable employment, but only in certain sectors, the outcome shifts from a general improvement to a kind of reconfiguration of labor inequality.
This shift in focus is very important. If technology does not destroy employment, but does transform its quality, then public policies should focus on facilitating the transition to digital sectors, improving training and skills, and reducing segmentation between sectors. It matters how many jobs are created, but also who gains access to the best ones.
“The effects of automation on employment are heterogeneous and depend on the type of technology, the institutional context and the time frame considered”
The relationship between digitalization and the labor market should be interpreted as a postpandemic portrait, not as permanent trends. The Eurostat statistics themselves, from which the analysis begins, have a limited frequency. In that uncertainty lies one of the major challenges of research in this field. Together with professors Iñaki Aliende and Lluís Franco, in our systematic review of thirty years of debate on the “end of work”, we found that the effects of automation on employment are heterogeneous and depend on the type of technology, the institutional context, and the time frame considered.
It is necessary to build monitoring systems that incorporate indicators of job quality, wage distribution, rate of technological adoption, and sectoral composition of work, in order to detect in time if digitalization begins to operate differently from how it has done thus far. The pattern identified in this analysis, the transformation of job quality more than the net destruction of jobs, may be reinforced, moderated, or reversed as artificial intelligence broadens its reach into cognitive tasks that, to date, are unknown.
The content of this article is the result of a study conducted by professors Iñaki Aliende (Complutense University of Madrid), Lluís Franco Sala (University of Barcelona) and Víctor Martín Cerdeño (Complutense University of Madrid) and that will be presented at the VII Economic Policy Meetings 2026 to be held at the University of Alicante on June 4 and 5, 2026.
The content of this article is the result of a study conducted by professors Iñaki Aliende (Complutense University of Madrid), Lluís Franco Sala (University of Barcelona) and Víctor Martín Cerdeño (Complutense University of Madrid) and that will be presented at the VII Economic Policy Meetings 2026 to be held at the University of Alicante on June 4 and 5, 2026.
