A third of people in the EU used generative AI tools overall in 2025, and fewer than half of them used them for work purposes, bringing the share to 15%. AI use at work varies widely across Europe, with clear regional patterns.
Artificial intelligence is becoming embedded in everyday life — from education to the workplace. Yet while many people use generative AI tools, far fewer apply them professionally.
ADVERTISEMENT
ADVERTISEMENT
In 2025, 15% of people aged 16 to 74 used AI for work, according to Eurostat. But that headline figure masks stark differences between countries.
Across 33 European nations, workplace AI use ranges from just 1.3% in Hungary to 35.4% in Norway, with Switzerland close behind at 34.4%.
“Norway’s higher share is entirely consistent with a strong digital public sector, high public trust, strong skills, and mature employer practices,” Prof Aleksandra Przegalińska from Kozminski University told Euronews Business.
Other high adopters include Malta (29.6%), Denmark (27.2%), the Netherlands (26.6%), Estonia (25.1%) and Finland (25.1%).
At the other end of the scale, fewer than one in ten people use AI at work in countries including Romania, Turkey, Serbia and Italy.
“The key point is that adoption is not just about enthusiasm: it follows investment in skills and workflows, plus trust-building governance that turns experimentation into legitimate, routine practice,” Przegalińska said.
A clear regional divide
The data reveals a strong geographic pattern.
Northern and Western European countries lead in adoption, while Southern Europe shows a mixed picture. Eastern and Southeastern Europe generally lags behind.
Among the EU’s largest economies, France records the highest workplace use at 18.4%, followed by Spain (17.9%). Germany is slightly above the EU average at 15.8%, while Italy is significantly lower at 8%.
Several smaller economies — including Luxembourg, Cyprus, Austria, Sweden and Belgium — also report relatively high usage, with rates between 20% and 25%.
Importantly, the figures reflect individuals using AI at work, not the share of companies adopting the technology.
A gap between personal and professional use
Looking at AI use overall and for work, there is a significant gap. Overall AI use in the EU stands at 32.7%, while use for work is 15.1%. This means that fewer than half of AI users, around 46%, apply it for work.
The gap varies significantly by country.
In Switzerland, Malta, Norway and the Netherlands, most AI users also use it at work. In contrast, countries such as Hungary, Romania and Serbia have much lower rates of workplace use among AI users.
Przegalińska said these differences come down to a mix of “capability” and “permission”.
Skills, structure and workplace culture
Capability includes digital skills, the share of knowledge-based jobs, and digital infrastructure such as broadband and cloud access.
Permission, meanwhile, is shaped by organisational culture and rules.
“Where employers provide approved tools, clear guidelines, and training tend to see faster uptake because employees feel safe using GenAI and know what is allowed,” she said.
OECD data shows that individual use of generative AI is rising rapidly, increasing by 68% between 2024 and 2025 in EU countries with available data.
“Businesses are using more AI too, and generative AI is a key driver of this surge,” said Nils Adriansson, economist-statistician at the OECD.
He stated that large firms are typically early adopters and have more opportunities to implement new technologies given their breadth of activities and resources.
The role of economic structure
The make-up of national economies also plays a role.
“The differences in the data may be explained by the different composition of national economies, with some countries having more industries and sectors where generative AI could more easily be deployed, such as knowledge and media work, ICT, research and development,” Prof Valerio De Stefano from York University in Toronto, told Euronews Business.
He added that some workers may underestimate how much they already rely on AI, as many commonly used tools are powered by it.
The data was collected in 2025, before the more recent spread of AI agents across the economy, suggesting that adoption rates could rise further in the near future.
