top of page
  • Facebook
  • Instagram
  • X
  • Youtube
  • LinkedIn
Search

From Machines to Minds: How AI is Reshaping Work and Income Distribution

A widely debated issue surrounding AI is that, as its use and diffusion throughout the economy increase, it becomes clearer which categories of individuals are likely to gain or lose from this new technological wave. Since the time of the Luddites, it has been common to distinguish the winners and losers of technological change between workers and capitalists. However, this distinction no longer seems adequate for several reasons. First, today’s workforce is highly heterogeneous in terms of skills, competencies, and the tasks performed. Second, over the past half-century, technological change has created winners and losers even within the labour force, favouring the wages and employment conditions of white-collar, high-skilled workers who were better positioned to use computers, software, and other digital tools.


With the advent of the latest wave of digital technologies, this pattern may be shifting again. Although these technologies share common technical foundations, their applications differ profoundly. AI represents the latest chapter in a long history of technological change that began with machines replacing muscle power and continued with robots automating routine tasks. But AI is distinct: it can also perform cognitive tasks traditionally undertaken by white-collar and professional workers.

Industrial robots and automated machinery have long threatened the jobs of low-skilled workers by taking over physical tasks such as moving, cutting, and packaging. Today, however, even white-collar employees and professionals engaged in routine cognitive tasks—including scientists involved in designing new products or production processes— face displacement risks from AI systems and intelligent software.


A central question, therefore, is how digital technologies are reshaping the distribution of income across production factors, especially given the strong geographical concentration of AI development in a few technological hubs. If AI creates new winners and losers by widening income differentials, these effects are likely to be particularly pronounced at the local level. As a result, the rapid diffusion of AI may further deepen regional economic disparities.


In a new paper, co-authored with Antonio Minniti and Klaus Pretter, we study this mechanism across Europe, examining the dynamics of the distribution of regional income between workers and capitalists.

We used data for 273 European regions (including the UK) between 2000 and 2017 to answer the following question: when a region becomes more active in AI innovation, does the share of income going to workers (the so-called labor share of income) change? We found several interesting results.

First, we find that although the distribution of income between workers and capital follows a national pattern, their dynamics over time are highly geographically differentiated within Europe (see Figure 1).


ree

Second, there is clear evidence that AI innovation in Europe is highly concentrated. A handful of regions, in countries like Germany, France, Sweden, and the UK, account for most AI patents. Southern and Eastern European regions lag behind (see Figure 2).


ree

Our inference illustrates that for every doubling of AI innovation, the labor share in a region drops by roughly 0.5% to 1.6%. To date, this implies an average reduction of nearly half a percentage point from an initial labor-income share of around 52%. While this effect is modest, it is not negligible—especially given that AI technologies remain in the early stages of their product cycle and our analysis does not yet capture the rapid diffusion of generative AI tools (e.g., ChatGPT, Claude, DeepSeek). We also find that, for high- and medium-skill workers (engineers, managers, professionals), the decline in labor share is driven primarily by wage compression, as their occupational structures have remained relatively stable. For low-skill workers, the pattern is more heterogeneous, indicating a mix of wage and employment adjustments.


References

 
 
bottom of page