Paul's POV
←All writing
AI Journey//5 min read

30% of EU Workers Use AI. That Number Hides the Real Gap.

The EU says almost a third of workers now use AI. Look closer and most of them tried it, a few rely on it, and the gap runs by job and company size.

30% of EU Workers Use AI. That Number Hides the Real Gap.

The European Commission says AI tools are now part of everyday work for around 30% of EU workers. It sounds like a tipping point. When I looked at where the number comes from, it told a different story, and one that matches what I see around me.

Quick Answer: The 30% comes from a large EU survey of more than 70,000 workers, and it counts anyone who used an AI tool for work at least once in the previous twelve months, mostly chatbots for writing and translation. A 2026 Commission survey puts regular use at work closer to one in four. Use is uneven: around 40% of white-collar workers use AI, against 6% in elementary jobs, and large companies are almost three times as likely to use it as the average EU firm. The real gap is between trying AI and building it into how you work.

01.What the 30% Actually Measures

The figure goes back to the AIM-WORK survey by the Commission's Joint Research Centre, which asked 70,316 workers in all 27 member states about AI and digital tools at work. Around 30% said they had used an AI-powered tool for work in the previous year. Most of that use was chatbots, and the most common tasks were writing (65% of users) and translation (59%).

That is real progress, but "used at least once in a year" is a low bar. It includes the person who asked a chatbot to polish one email in March and never went back.

Other figures point the same way. A Commission survey from February and March 2026, presented at the Brussels Economic Forum, found that slightly more than half of Europeans use AI in some form, but only about one in four use it in their job. Those who do reported saving 7.4 hours a month on average. That is almost a working day a month for the people who use it well, and nothing for everyone else.

02.Where the Gap Runs

The Commission's own summary from 23 October is more interesting than the headline. It says many jobs that could use AI don't, especially in smaller companies and among mid-career staff.

  • By job: around 40% of white-collar workers use AI at work, against 6% of workers in elementary jobs.
  • By company size: Eurostat counts 20% of EU companies with ten or more employees using AI in 2025, up from 13.5% a year earlier. Among large companies it is 55%.
  • By country: the share of companies using AI ranges from 42% in Denmark to about 5% in Romania.

A March 2026 study by economists Bick and co-authors, published by Brookings, compared AI adoption in the US and Europe. It linked the gaps partly to who the workers are and what kind of companies they work for, and partly to something simpler: whether employers actively encourage people to use AI.

03.What I Recognise in This

I wrote about the AI adoption paradox earlier this year: most people who try AI tools quietly stop within weeks, because nobody teaches the in-between skills that make the tools useful. The EU numbers look like the European version of the same story.

Around me, at work and outside it, many people use AI as a more advanced Google search. They ask a question, read the answer and move on. That is useful, but it leaves most of the advantage on the table. In my view, the gap between people who really work with AI and people who use it as a smarter search engine is now getting wider.

For me the change came from using AI on real work every day, not from trying it. A year ago I had never built software. Since November 2025 I have started six projects with AI coding tools, mostly in the mornings and evenings. None of that came from a course. It came from repetition, and from learning where the tools are reliable and where they are not.

The mid-career point worries me most. People in the middle of their careers carry most of the judgement in an organisation. If they are the ones least likely to build AI into their work, the gap will not close by itself, and it will show up later as the experience gap between juniors who grew up with AI and seniors who never really adopted it.

04.What Would Close It

The Commission's answer is faster and more targeted training. That helps, but the research suggests three simpler things matter as much.

  1. Encouragement from the top. People use AI more when their employer clearly says it is expected and allowed.
  2. Real work, not demos. Training on your own tasks beats a general course on prompts.
  3. Time to practise. Seven hours saved a month only arrives after a few weeks of being slower.

The Commission closes its European AI Innovation Month with the Apply AI Summit in Brussels on 17 November. I hope the conversation there moves from how many people have access to AI to how many actually work differently because of it. That is the number I would like to see next year.

Cover image generated with Krea AI.

Sources

More writing →