ECB Blog: Current State of AI Adoption and Employment in Europe - Some Thoughts

I. The Current State of Employment Based on Empirical Data

Concerns that AI will replace human labor and lead to massive job losses arise repeatedly. In the United States, large-scale layoffs of thousands of employees have been reported at companies like Amazon and Target, with AI cited as one of the contributing factors. However, at least in Europe at the present time, the data points in the opposite direction.

On March 4, 2026, Laura Lebastard and David Sondermann of the European Central Bank published a report on the ECB blog analyzing the relationship between AI adoption/investment and hiring behavior in European firms [Note 1]. The data used was from the Survey on the Access to Finance of Enterprises (SAFE) conducted by the ECB, covering 5,000 companies. As a prerequisite for considering the impact of new technology adoption on legal systems and society, it is essential to scrutinize objective data from the real economy. This article summarizes the findings of that report and adds some considerations.

1. Expansion of Use and Limited Direct Investment

While AI usage among European firms is expanding rapidly, instances involving direct investment in the technology remain limited. Two-thirds of the surveyed companies report AI usage by employees, with the usage rate reaching nearly 90% among large enterprises with over 250 employees. Even among small firms with fewer than 10 employees, about 60% use it, meaning the difference in scale is not as large as one might imagine.

In contrast, only about one-quarter of all firms are actually investing in AI technology. It is understood that the proliferation of online tools allows companies to utilize AI without significant capital expenditure, and this ease of entry supports widespread use among small firms.

2. Impact of Intensive Use on Employment

When comparing firms simply by whether or not they use AI, no statistically significant difference in hiring or firing behavior is confirmed. However, the picture changes when distinguished by the degree of usage. Firms that use AI intensively have a roughly 4% higher probability of hiring additional staff compared to those that do not. Firms that invest in AI also have a nearly 2% higher probability of hiring compared to those that do not invest.

This increase in hiring is led by small firms, while in large enterprises, the impact of AI on employment acts neutrally. It is speculated that highly skilled personnel are newly required in the process of operating and integrating new technology into business, but the attributes of the hired personnel cannot be identified from the survey data alone.

II. The Future of Employment Divided by Purpose of Introduction

1. Differences Between R&D and Labor Cost Reduction

Driving employment growth are groups of firms that utilize AI for the purposes of research and development or business expansion. Here, we see a direction of aiming to develop and introduce AI technology while maintaining existing production processes, or attempting to scale up more rapidly through AI.

On the other hand, in firms that introduce AI primarily for the purpose of reducing labor costs, a negative impact on hiring and a positive impact on layoffs are observed. It is true that there is a movement in some areas to replace human labor with machines. However, only 15% of firms using AI cited labor cost reduction as the main reason, and at present, this has not reached the point of offsetting the net positive employment effect observed overall.

2. Short-term Expectations and Long-term Uncertainty

Even in hiring plans looking one year ahead, firms specifically planning AI investment tend to show more positive expectations for employment growth than firms without such plans. This trend is observed regardless of the scale of the planned AI investment, and at least in the near future, the prospect that AI investment will be a factor in suppressing hiring is not supported at this time.

However, it is necessary to carefully determine whether these findings hold true over different time horizons. A survey by the ifo Institute suggests that many German companies expect layoffs to occur due to AI introduction over a medium- to long-term horizon of five years [Note 1]. As the report itself acknowledges, the current positive impact on employment is due to the fact that AI has not yet significantly transformed corporate production processes. If business processes are comprehensively reviewed as technology matures, a phase may arrive where the substitution effect on employment outweighs the creation effect.

III. In Conclusion

The current situation in Europe as shown by the ECB report is that in the initial stages of technology adoption, AI is functioning as a complementary means of business expansion rather than a substitute for workers. At present, firms whose primary goal is labor cost reduction remain in the minority, and many companies are newly seeking personnel who can utilize and manage AI.

The scale of investment in and the timing of AI adoption differ significantly between the United States and Europe, and it is difficult to determine the future of one from the short-term data of the other. Furthermore, even within Europe, this optimistic short-term data does not guarantee a slow pace of structural change. When technology reaches a stage where it further increases its autonomy and significantly reshapes the nature of work, how legal systems will reconcile the protection of workers with the demands of technology utilization is an issue that should be continuously questioned from now on. I intend to continue observing the reality where technological development and labor market transformation intersect. For details, please refer to the original source.

(Magazine) "Random Thoughts on AI and Law"

*Please refer to the following for the table of contents

note General Terms and Conditions Article 3, Paragraph 2, First Sentence
3.2 The copyright of digital content created by the creator belongs to the creator.
P.S. This is shared for academic discussion only.