European companies that use artificial intelligence intensively have a 4% higher probability of hiring than those that use it occasionally, according to an analysis by the European Central Bank covering 5,000 companies in the eurozone. This empirical reality questions the dominant discourse on automation as a job killer and reveals a paradox: while experts debate future job losses, companies investing in AI are recruiting more.
The CEPR study simultaneously confirms that AI adoption increases labor productivity by 4% on average in the EU, without job reduction in the short term. These data reframe the debate: the issue is no longer protecting against a phantom threat, but understanding why these productivity gains benefit European companies unequally.
The essentials
- Companies using AI intensively have a 4% higher probability of additional hiring; those investing in AI have 2% more
- AI adoption generates 4% productivity gains without short-term job destruction according to CEPR
- Productivity gains are concentrated in medium and large companies, accentuating inequalities
- More than a quarter of German companies anticipate job cuts within five years according to the ifo Institute
Companies that adopt AI hire more than they lay off
Two ECB economists, Laura Lebastard and David Sondermann, analyzed recruitment behaviors of 5,300 companies in the eurozone. Their method controls for company size, age, economic prospects, revenue, profitability, and neutralizes sectoral and geographic effects.
The main finding contradicts fears: companies that make significant use of AI have approximately 4% higher chances of hiring additional personnel. Those investing in AI have nearly 2% higher chances of recruiting.
This positive dynamic is explained by the strategic use of technology. Companies that mobilize AI for R&D and innovation are those that contribute most to overall employment growth. Companies that hire thanks to AI are those that use it to develop, not simply to save costs.
Conversely, companies that use AI primarily to reduce labor costs experience negative effects on hiring. This minority represents approximately 15% of organizations using these technologies.
4% productivity gains without job destruction
The Centre for Economic Policy Research study provides causal validation of these observations. The analysis examines how AI adoption affects productivity and employment in more than 12,000 European companies.
AI adoption causally increases labor productivity levels by 4% on average in the EU. This effect is statistically robust and economically significant, although more moderate than the transformation scenarios predicted by some observers.
Crucial for the employment debate: we find no evidence that AI reduces employment in the short term. The absence of negative effects on employment, combined with significant productivity gains, reveals a specific mechanism: AI increases worker output without displacing labor.
This complementarity translates into wages. Workers in companies that adopt AI benefit from higher wages, both at the aggregate level and per employee.
Inequalities between companies are widening
This optimism masks a more contrasted reality: productivity gains are distributed unequally. Medium and large companies, as well as those with the capacity to integrate AI through investments in intangible assets and human capital, record significantly larger productivity gains.
This digital divide is measurable. In France, 9% of companies with fewer than 50 employees use AI, 15% of companies with 50 to 249 employees, and 33% of those with 250 or more employees. The gap between the largest and smallest companies is growing, increasing from 16 to 24 percentage points between 2023 and 2024.
Competitive advantage accumulates. Companies using AI concentrated 49% of total revenue in 2024 and 40% of total employment, compared to 40% and 31% respectively in 2023.
This geographic concentration reinforces disparities. Financially developed EU countries, such as Sweden and the Netherlands, match US adoption rates, with approximately 36% of companies using big data analytics and AI in 2024. In contrast, companies in the financially less developed EU economies, such as Romania and Bulgaria, significantly lag behind, with adoption rates of approximately 28% in 2024.
The long-term horizon remains uncertain
This short-term analysis does not close the debate. The ifo Institute published a survey showing that many German companies anticipate job cuts related to AI over a five-year horizon. More than a quarter of German companies expect AI to lead to job losses over the next five years.
The ECB confirms it: AI has not yet fundamentally transformed production processes. That moment will come. Christine Lagarde emphasized that massive investments in AI were translating into productivity improvements, but that consequences for the labor market were not yet visible.
The temporal paradox is striking. While Europe debates preventive policies against technological unemployment, the current positive effect is driven by small companies, while AI remains neutral on employment in large structures in the short term.
This transition window reveals the true issue: not protecting against an employment apocalypse, but ensuring that productivity gains benefit the largest number of companies before the transformation of processes reshuffles the cards.