Why 57% of Employees Hide Their AI Use (And What It Means for HR)

While you've been debating AI policy, your employees have been uploading company secrets.

Jun 17, 2025

Welcome to FullStack HR, where we talk about the future of work.
If you haven’t subscribed, join the 9500+ smart, curious folks by subscribing here:

Happy Tuesday!

When I launched FullStack HR in 2022, well before the AI hype, I set a long-term goal of 10,000 subscribers. I even bought a special bottle of champagne to open if we reached that number before 2030.

We are now very close.

If you find this newsletter useful, please help me open that bottle by sharing it. If posting on LinkedIn feels awkward, share it with your team on Teams or Slack instead.

Thanks in advance for spreading the word! :)

On to the update.

Right now, in your building, someone is asking ChatGPT to write a customer email using proprietary pricing data. Someone else is having Claude review a confidential contract. A third person just automated a workflow that could eliminate two jobs, including their own.

Did you think I made up all of the above?
Here’s more stats for you.
(Sources in the bottom of the article.)

Shadow AI is everywhere

57 % of 48 340 employees hide their AI use from managers.
32 % of office workers keep AI shortcuts secret; 30 % do so because they fear layoffs.
70 % rely on free public tools; only 42 % use employer-approved ones.
48 % have uploaded sensitive data into public chatbots.
66 % trust bot answers without checking; 56 % have made AI-driven mistakes.
Only 47 % have had any AI training.

Adoption outruns policy

78 % of firms now use AI in at least one function; 71 % already use generative AI.
Responsible-AI roadblocks: 51 % cite knowledge gaps, 45 % budget limits, 40 % regulatory uncertainty.
Where Responsible-AI policies exist, 42 % report better efficiency and lower costs; 34 % see higher customer trust.

Leadership signals

82 % of leaders say 2025 is the year to rethink strategy around AI.
81 % expect AI agents to be woven into core workflows within 12–18 months.
Only 24 % have rolled AI out company-wide.

Money floods in

Generative-AI private funding hit $33.9 billion in 2024.
VCs poured $82 billion into North-American startups in Q1 2025, with AI the main driver.
Generative-AI deals took 57.9 % of global venture dollars that quarter.
By late May 2025, U.S. Gen-AI funding alone had already cleared the $50 billion mark.

Why all this data?
Because it shows one clear thing.

The blocker for AI adoption is not tech.

It’s people.
It’s organization.

Your skills as an HR professional matter more than ever before.
We need you.
We need your skills.

We need more HR-people stepping into where IT sits and loudly scream “we’re her now”.

Don’t wait. Do it.
Do it now.

You have the right skills.
Make use of them.

Sources

(I’ve been gathering AI-reports for the last 2,5 years in a folder in Google Drive. I gace ChatGPT access to the folder and did a Deep Research on the investment part, I deliberately have been quite conservative with the dollar amount, in reality it’s probably higher. But my number is stil insanely high.)

Roi Ezra

Jun 20

So well put. The real friction isn’t in the tools, it’s in trust. I’ve been writing about this gap too: when teams hide what works because the system doesn’t feel safe, it’s not innovation we lack, it’s permission. There’s so much to learn when we shift from enforcement to invitation.

Expand full comment