Without management, artificial intelligence does not produce value
The adoption of Ai is not only a technological issue but a matter of corporate governance and managerial responsibility
Giuliano Noci's article published in the pages of Il Sole 24 Ore on 11 April captures an essential point: artificial intelligence is helping to further widen the productivity gap between the United States and Europe, which has been in place for over thirty years. As Federmanager, we believe that this new wave of technology represents a test case for the country's political and economic class: the adoption of Ai is not just a technological issue but one of corporate governance and managerial responsibility.
The evidence is clear and, in some respects, worrying. The study Mind the Gap: AI Adoption in Europe and the US published in March 2026 by the Brookings Institution is the largest comparative survey ever conducted on AI adoption in the workplace, with more than 55,000 workers interviewed in seven countries, and captures a widening gap.
In 2026, 43% of US workers will use generative AI in their work. In Europe the average drops to 32%, while Italia stops at 26%, last among the countries considered. But the gap is not only quantitative: in the US, AI is used for 5.2% of working hours, more than three times as much as in Italia, which stands at 1.6%. This differential already translates into a significant productivity advantage for the US, up to 1.3 percentage points. Between 2022 and 2024, European sectors with higher AI adoption rates experienced 2-5 percentage points higher cumulative productivity growth than lagging sectors.
The central question, however, is not 'how much', but 'why'. Why, given equal access to the same technology, are the results so different between companies in Reggio Emilia and those in Houston?
Brookings' answer is precise and of great relevance for economic policy as well. Only 55% of the gap between the US and Europe is explained by 'traditional' factors: sectoral composition, company size, demographic structure of the workforce. The remaining 45% depends on a single factor: whether or not one's employer encourages the use of AI. So, it is not a question of mere availability of tools. Not of training. Not of public subsidies. But of managerial input on what is expected, what is valued, what is safe to experiment with and use in the company. This is what enables workers to adopt AI productively.

