In this guest blog for International HR Day 2026, Aco Momcilovic and Raffaele Credidio of the EAPM AI Working Group explore what the data is telling us about AI adoption in HR teams across Europe, and where people professionals need greater support.
There is a scene in countless old movies where someone receives terrible news and, in the next shot, is inexplicably at a dinner party smiling. That is roughly what it feels like to look at how organisations talk about AI in HR versus what our data shows is actually happening.
AI adoption by firms has more than doubled across OECD economies in just two years. SHRM’s 2026 State of AI in HR report finds that 87% of CHROs forecast greater AI adoption in HR this year. Everyone, apparently, is moving fast. So when the EAPM AI Working Group surveyed 327 European HR professionals — across 30 countries, the majority Heads of HR with over 15 years of experience — we were not quite prepared for what came back.
It turns out the dinner party is largely a fiction. The gap between the headline and the reality is where the most urgent story actually lives.
43% of HR departments in our survey use AI informally — individuals experimenting without structure, policy or backing. Only 19% describe AI as genuinely integrated into workflows. And 66% say their organisation has no clear guidelines for AI use in HR. These are not small organisations. These are experienced senior HR leaders — and the picture they paint is one we started calling “sanctioned improvisation”: AI is allowed, it just isn’t managed.
This connects to a warning issued back in 2023 by Peter Cappelli and Nikolai Rogovsky in an ILO Working Paper: while firms enthusiastically embrace AI in HR, their understanding of how it affects the workforce often lags behind or is not viewed as a priority. Three years on, across 30 European countries, the warning has gone largely unheeded.
64% of HR professionals we surveyed have received no formal AI training — no course, no workshop, nothing. 35% rely entirely on self-learning. Only 15% have mandatory, structured programmes. And 66% told us their organisation does not systematically invest in AI competency development for HR. That is not a gap. It is a choice.
The paradox is this: 60% of respondents say they understand the limitations of AI — bias, hallucinations, lack of transparency. It was the only question in our survey where a clear majority agreed. HR professionals know what can go wrong. But knowing the risks without the governance structures or organisational support to manage them is not preparedness. It is awareness without agency.
89% of HR professionals told us that responsibility for HR decisions must always remain with humans — the strongest consensus in our entire survey. 61% said that when AI contributes to a wrong decision, the HR professional should be personally responsible. And 71% said accountability for AI-related errors is not clearly defined in their organisation.
Read those three together. HR professionals are expected to carry personal accountability for AI-influenced decisions — in organisations that have never defined what that accountability means. No guidelines. No training. No safety net.
This is not abstract. The EU AI Act is in its critical implementation phase, and many HR AI tools will be classified as “high risk,” triggering mandatory human oversight requirements. Organisations without defined accountability are not just culturally unprepared. They may soon be legally exposed.
European AI adoption sits at around 20% on average — but this conceals enormous variation. CEPR research found that European workers use AI for roughly one-third the hours of their US counterparts, with country-level rates ranging from 26% in Italy to 36% in the UK. The challenge is not uniform. An HR manager in Stockholm is navigating a very different landscape than one in Bucharest, and governance gaps will look different in each context.
The area where the gap between aspiration and reality is most striking is talent management. 56% believe AI will strengthen HR’s strategic role long-term. Yet 60% use no AI in talent management whatsoever — not for identifying high-potential employees, succession planning or career pathing. The most consequential HR decisions remain almost entirely unassisted by AI in 2026. Organisations are not failing here because they’ve tried and it hasn’t worked. They are failing because they haven’t tried at all.
The HR professionals in our survey are not resistant to AI. They understand the risks better than most. They want more AI in the future. They are simply operating without training, without guidelines, without budget in many cases — and being asked to be personally responsible for outcomes in that environment.
The ILO paper put it well: the choice between high-road and low-road approaches to technology adoption is not about the technology. It’s about whether employers invest in people alongside systems. For most European HR departments in 2026, the answer has been: not yet.
International HR Day feels like the right moment to say this plainly. The profession is ready. The organisations aren’t.