Official sources checked 24 October 2026. The Commission summary was published at 15:59 CEST on 23 October, inside the coverage window; the underlying JRC brief was released earlier. The summary says about 30% of EU enterprises adopted AI, while the JRC brief and Eurostat report 20% in 2025. AISFY uses the primary-report figure and does not silently reconcile the discrepancy. Sources: European Commission DG Employment — AI and skills in Europe; European Commission Joint Research Centre — AI and Skills; Eurostat — 20% of EU enterprises use AI technologies.
23 October 2026, global workforce-learning signal: the European Commission published a new summary of Joint Research Centre (JRC) evidence on AI exposure, adoption and skills. Its most useful message for COOs, CHROs, CAIOs, HR and L&D leaders is that exposure does not determine one training response. Leaders need to examine the work itself—especially judgment, accountability and human interaction—before deciding whether people need upskilling within their current roles or support to reskill toward different work.
What does the JRC evidence say about AI exposure and adoption?
The Commission summary published on 23 October highlights an adoption gap: many jobs contain tasks that AI could assist, but use remains uneven. The full 21-page JRC brief reports that 30% of EU workers used AI at work in 2024–2025, ranging from close to 40% among white-collar workers to 6% in elementary occupations. It also reports that 20% of EU enterprises with at least ten employees used an AI technology in 2025, up from 13.5% in 2024.
The Commission news page describes enterprise adoption as about 30%, but the JRC brief and the underlying Eurostat release report 20%. We use the primary-report figure. More broadly, the JRC combines seven occupational-exposure measures with EU labour-force, worker-use and enterprise-adoption evidence. Exposure measures technical overlap between job tasks and AI capabilities; it does not show that a worker has adopted AI, that a task will be automated, or that employment effects have occurred.
Which roles need upskilling, and which may need reskilling?
The brief pairs exposure with human-AI complementarity: the degree to which communication, responsibility, physical context, judgment and required training keep people central to the work. It places medical practitioners, veterinarians and professional-services managers among examples with high exposure and high complementarity, where the case is stronger for learning how to apply AI within the occupation. General office clerks, keyboard operators and some finance roles combine high exposure with lower complementarity, which raises a stronger case for reskilling across occupations.
This is a planning distinction, not a forecast of job losses. The JRC says evidence through 2025 does not show economy-wide displacement and that occupational labour-market signals remain too unclear for precise reskilling pathways. For a training buyer, the implication is to segment by task and role rather than assigning one course to everyone. A role with decision accountability may need practice in verification, escalation and process redesign; a role with a larger share of substitutable tasks may need broader transition support.
Why do short, role-relevant learning formats matter?
The JRC's 2025–2026 mapping of English-taught programmes listed by StudyPortals found AI content in 10.6% of EU master's programmes and 7.1% of short courses; the comparable US short-course share was 10.5%. The authors describe the mapping as supply-side evidence, not proof of course quality, participation or learning outcomes. They also note that workers aged 25 to 44 held the highest shares of employment in the two most exposed occupational groups, making flexible provision especially relevant to people already in work.
The brief argues that workforce AI literacy should combine conceptual understanding, practical application, critical evaluation and human-centred decision-making. It says initiatives are more likely to be useful when they follow a capability assessment, connect to concrete organizational objectives and include implementation and evaluation. The findings concern the EU and do not establish UAE adoption levels, training supply or workforce outcomes. UAE buyers can use the decision logic, but still need local role evidence and their own baseline.
AISFY reflection: start with the role, not the course catalogue
AISFY reflection: The decision is not simply who is “exposed to AI.” Ask which tasks may change, where human judgment and accountability remain essential, what people can already do, and whether the intended pathway is better performance in the current role or preparation for different work. That assessment should determine the practice tasks, facilitator expertise, manager support and evidence collected after learning.
Begin with an AI training needs analysis, compare role-based team learning choices, and define how workshop learning will be evaluated before selecting a programme. If the roles and tasks are already clear, use them to discuss an enterprise AI workshop.
