Two-Thirds of European HR Teams Are Using AI — Are They Doing It Responsibly?

If your HR team hasn't touched AI yet, you're now the exception. 

Our latest research shows 67% of departments have already gone from testing the waters to using it every day. From candidate screening to onboarding chatbots and real-time performance insights, AI now touches a lot of HR workflows.

Here's the catch. Adoption and governance aren't moving at the same speed. Only 36% of HR leaders say they're ‘very confident’ they could actually defend their AI use if a regulator, employee or journalist came calling.

That means most teams have adopted AI much faster than the guardrails to match. And with strict AI regulations landing across Europe this year and next, the countdown to getting governance in order is already ticking.

Where is the confidence gap coming from?  

Ask any HR leader if they have AI safeguards in place and 96% will say yes. Push a little further to ask what those safeguards are and the picture gets patchier fast. 

Only 51% have someone actively reviewing AI-assisted decisions, while just 36% have completed and documented bias testing.

Most safeguards start and end with a policy document, not a hard look at the workflows and outputs AI is actually generating. That gap matters more now that EU AI Act obligations and GDPR are the ones asking the questions, not just your own compliance team.

It's also why HR leaders are drawing such firm lines around what AI doesn't get to touch. 92% say there's at least one decision AI should never make on its own. Terminations and layoffs top the list (45%), followed by disciplinary action (40%) and pay decisions (37%). 

German HR teams are the most emphatic of all: 61% say AI should never be solely responsible for a termination, shaped by the country's co-determination laws, which give works councils a legal seat at that table.

The bottom line? HR trusts AI for the parts of the job that were never really about judgment, but it’s starting to draw a hard line around the parts that are.

What's slowing everyone down?

Data security and privacy is the single biggest barrier to broader AI adoption in HR, cited by 41% of leaders. That’s more than budget, lack of internal expertise and skepticism about whether the tools even work.

Regulatory uncertainty comes next at 31%, tied closely with the headache of getting AI to connect with existing HR systems and concerns about bias creeping into decisions, both at 30%.

These are no longer abstract concerns, especially for European businesses. Between the EU AI Act's incoming high-risk provisions and existing GDPR obligations, there’s a lot less room to wing it. Canada's AIDA and similar regimes are doing the same elsewhere. 

Australia offers an interesting contrast. Its National AI Plan actively encourages adoption, treating AI as an opportunity rather than a risk to contain. That lighter-touch approach may be exactly why Australian teams report the highest rates of advanced adoption at 39%, nearly double Europe's 21%.

In any case, the departments making the most progress aren't waiting for the regulatory picture to get simpler. They’re investing in better tools to monitor AI compliant workflows and more training to get their teams up to speed.

What ‘responsible usage’ actually looks like

Not more paperwork, for a start. 

The teams reporting the highest confidence in their AI compliance (79% among advanced adopters versus just 36% overall) aren't the ones with the thickest binder of policies. They've audited which AI tools touch which decisions and made someone in HR or legal responsible for it – and then gone back through vendor contracts to check the right documentation is there.

That's the pattern worth copying. Confidence comes from being able to point to what was checked, who checked it and when. Not just having a policy that says AI is being used responsibly. 

It also doesn't mean pulling back on AI itself. Only 8% of HR leaders think AI could handle most decisions independently even with strong safeguards in place. But that number climbs to 14% among advanced adopters and UK teams, the same groups already showing the strongest governance habits. Which confirms the idea that maturity and comfort with AI should grow together, not apart.

Would your AI decisions survive a closer look?

Here's the test that matters. Not whether your HR team uses AI, but what happens when someone asks to see the receipts. Some HR teams could pull that up tomorrow. Most would need a few days and a fair amount of scrambling to even figure out what to present.

It’s a lot easier when the system running your AI workflows already keeps a paper trail without you asking it to. A platform like Rippling, with permissions and activity logs baked in from day one, means the answer to what was decided (and why) is already sitting there. It’s not something someone has to reconstruct after the fact.

That's really the story here. Two-thirds of HR teams have already made the call on AI. What separates the ones doing it responsibly from the rest is whether they can back that call up.

Make sure to download the full report below to get the full picture from over 1,000 HR leaders, including country by country breakdowns for the UK, Germany and beyond.