It's Not the AI That Scares HR - It's the Data

Nobody expected HR to be the one setting the pace on AI. The numbers say otherwise.

Our latest survey of 1,000+ HR leaders across Europe, Canada and Australia found that 67% now rate themselves as intermediate or advanced AI users. The only function ahead of them is R&D – the team whose literal purpose is experimenting with new tech.

Scratch the surface a little further and you’ll find a contradiction: 92% of those same HR leaders say there's at least one decision they'd never hand to AI. It turns out HR wants the speed and scale AI promises, but it doesn't trust the tools (or the data behind them) enough to completely let go of the wheel. 

That tension is the real AI story in HR right now. It has less to do with AI itself than with the data infrastructure propping it up.

What day-to-day AI looks like for HR teams

Let’s dig deeper into that 67% stat to see what's actually happening on the ground in HR departments. First of all, it's rarely one big and flashy AI rollout; more likely a dozen smaller smart capabilities stacked on top of each other:

• Resume screening that used to take a recruiter half a day now takes twenty minutes.

• Policy drafts that went through three rounds of ‘can someone just write the first version’ now start with a draft already on the page.

• Scheduling conflicts, PTO questions, benefits FAQs and the stuff that used to clog an HR inbox all day, all increasingly handled before a human ever sees them.

So this isn't HR jumping on a trend. The number one reason leaders give for adopting AI is cutting down admin workloads so people can spend time on things that matter. And it's paying off. At companies growing more than 10% a year, 48% have reached advanced AI adoption, compared to just 9% at companies that aren't growing. 

That's the part worth sitting with, plus the fact 69% of HR leaders now call AI ‘very’ or ‘extremely’ important to their strategy over the next 12 months. It shows HR has moved beyond testing the waters and is already restructuring core workflows around these tools.

The line HR won't cross

For all the speed HR is picking up from AI and automated workflows, it isn't handing everything over.

We asked the same 1,000+ leaders where they draw the line and 92% say there's at least one call they'd never let AI make on its own. The most common examples were firing someone, disciplinary action, pay decisions or anything where a person's livelihood is on the line. 

Given how deep HR already is into AI, this level of caution stands out. Departments are already beginning to set clear boundaries for fair usage, partly fuelled by new, stricter AI regulations rolling out across Europe. 

Germany is furthest along with this. 61% of German HR leaders say AI should never be solely responsible for terminations or layoffs, the highest of any country in our study. Not because they trust AI less across the board, but because co-determination law already requires works councils to be part of decisions like these. 

For them, whether or not AI should be in the room for certain conversations isn't even a debate. The law settled it before the tool showed up.

Put the data points side by side and a pattern emerges. The more consequential the decision, the less HR wants AI anywhere near it, regulation or not. Which raises the question: if HR already knows where the line is, why does the idea of workplace AI still make people nervous?

Where is the fear coming from?

When you ask HR leaders what's actually holding them back from going further with AI, the top answer isn't ‘we don't trust the technology’ or ‘we don't have the budget.’ 

It’s data security and privacy, cited by 41% of leaders and ahead of every other barrier on the list, including lack of trust in AI decision-making (27%).

That reframes the entire conversation. HR isn't all that nervous about AI making bad calls. The underlying issue is around where the data lives, who can see it, whether it's protected and what happens if it ends up in the wrong hands.

And the uncertainty shows up in the numbers. Only 36% of HR teams say they're ‘very confident’ they can comply with new AI regulations, despite 46% already running data protection impact assessments and 51% reviewing vendor contracts for compliance documentation, just to keep pace.

That's the overhead of bolting AI onto infrastructure that was never built for it, and it multiplies with every tool you add to the stack.

Fix your foundations

Quiz most HR teams on where their data lives and the answer is ‘everywhere.’

It's usually scattered across a patchwork of systems never designed to talk to each other, each with its own logins, rules and blind spots. Every extra system in that patchwork is another access list to manage, another gap where an ex-employee's login still works and another place a breach can start.

A platform built around one unified data layer solves exactly that problem. Instead of a dozen vendors each holding a slice of employee data, everything sits in one place, under one set of access controls, with one clear audit trail of who touched what and when.

That's the fix. Give AI a single, permissioned, audit-ready place to work from, and trust stops being the thing standing between HR and the next stage of adoption.

Want the full picture? Download our report below for the full data from 1,000+ HR leaders on AI adoption, governance and where human judgment still leads.