Your employees are already using AI in secret - the question is whether your workplace makes it safer to admit that, or to hide it.

Hello Humans.

And welcome to the weekly podcast that aims to address an important AI development in no more than five minutes. My name is Barry Phillips.

Today: your staff are using AI. Why aren’t they telling you?

Imagine a colleague who finds a way to finish a two-hour task in twenty minutes. They could share it with their manager. Instead, they keep quiet.

Perhaps they expect a telling-off. Perhaps they’re wondering whether they’ve just demonstrated that their job could disappear.

That uncomfortable calculation is HR’s business.

New Deloitte research, published on the sixteenth of September, surveyed twenty-five thousand UK workers. Among those using generative AI for work, thirty-one per cent said they used it without their employer’s knowledge.
That’s almost one in three users. And there’s a name for it: shadow AI.

It means AI being used outside the organisation’s awareness or approved arrangements. 
Those are UK findings, but the question for HR teams in Ireland is just as relevant: do we actually know which tools our people are using?

Before we reach for the disciplinary policy, another finding deserves attention.

Sixty-four per cent of weekly users worried their managers would think AI could do their jobs.

Imagine the mixed message: “Please embrace AI. And do tell us exactly how much of your role it can replace.”
We shouldn’t be surprised if that conversation goes rather quiet.

But there are real risks here.

Imagine an HR adviser pasting a grievance statement into a personal AI account for a quick summary. That statement might contain medical information and allegations about colleagues.

Has anyone checked whether that tool is approved to handle those details?

So, here are three things I’d ask HR to do this week.

First, make it easier to tell the truth. Ask people what they’re using, what helps, and what worries them. Give them a clear way to raise concerns.

A useful opening question is: “What are people using that our policy hasn’t caught up with?”

Second, make the safe option practical. Agree approved tools with IT and your data protection colleagues. Explain what information can go into them. Demonstrate a realistic HR task, including checking the answer.

“Use AI responsibly” is not much help at half past four on a Friday.

Third, talk openly about time saved. Will it go towards better work, learning, or an ever-growing workload?

You don’t need to promise that jobs will never change. You do need to be honest about your plans.

This week, ask yourself: if somebody admitted they’d been using AI without approval, would our first response make the next person more likely to speak up, or less?

Because if honesty about AI feels more dangerous than using it in secret, your policy has already failed.

That’s it for this week. Until next week bye for now.