Summary

  • The European Commission put a new JRC study on AI skills in the spotlight on 23 October. Thirty per cent of EU workers already use AI at work, while the training offer favours long degrees over short courses.
  • In the JRC's catalogue of English-taught EU programmes, 10.6% of master's programmes contain AI content; the share among short courses is 7.1%. The distinction matters to workers whose jobs cannot be put on hold for a degree.

AI has found its way into the office. Its training plan is still looking for the lecture hall. A new Commission account of the Joint Research Centre's work says roughly three in ten EU workers use AI on the job. The original JRC study, released earlier in October, describes a much slower response in the kind of education that an employed person can fit around a working week.

The mismatch has numbers, but their denominators matter. The JRC examined a 2025–26 catalogue of English-taught programmes across the EU. AI appeared as a principal or secondary subject in 10.6% of master's programmes, compared with 7.1% of short courses. These are shares within each format, not a count of places available to any particular worker. The study's point is about design: much of the visible AI offer sits in degree programmes, while shorter, flexible routes to retraining remain comparatively thin. The people most exposed to AI are largely already employed and often 25 to 44 years old.

A master's timetable is a poor substitute for training that can reach them this quarter.

Nor is the worker pool uniform. The JRC's worker survey puts AI use close to 40% among white-collar employees and at 6% in elementary occupations. Its Eurostat-based measure finds 20% of EU enterprises with at least ten employees using an AI technology in 2025, up from 13.5% a year earlier. Worker use and company adoption count different things. Together they show why a single ceremonial course on “the future of AI” would be an impressive way to miss the people actually using it.

The governance problem begins where the training brochure ends. The Commission's account of AI Act Article 4 places measures to support AI literacy with system providers and deployers for staff and others operating AI on their behalf. The Commission and member states have supporting roles. The JRC lists the Union of Skills, the AI Continent Action Plan and an AI Skills Academy among the policy and funding machinery. The worker still needs a course that explains the system on their desk, in hours they can attend. A framework cannot take that shift for them.

Education has a second clock. The JRC reports AI-related competences in roughly a third of national school curricula and cites 29.7% of teachers taking professional development on AI for teaching and learning. Schools are preparing the next workforce while employers are already asking today's staff to exercise judgement around live tools. The two schedules meet only if institutions build a route between them.

There is no magic number of certificates to hit. The practical measure is whether a clerk, technician or manager can learn what their own system does, where its output fails and who decides when to use it. Europe has announced the destination. The JRC has measured the length of the walk from the workplace to the classroom.

Sources