Copy article

RAW - The Employer / Employee AI Disconnect - RAW

ended 16. August 2026

****NEWSPAGE RAW**** Unpublished raw News Alert

Recent ONS research (published here on 20/07/26) found that 55% of employed and self-employed respondents reported using AI for work or education, while approximately 35% of businesses with ten or more employees reported using AI technology (same report). 

A separate 2026 government survey (published here on 18/06/26) found that, among businesses using AI, only 17% had any policy or guidance governing it. Just 5% had a formal written policy. At the same time, 73% of businesses handling digital data said they would be uncomfortable with their information being used to train external AI models.

This raises a practical question: do employers know what staff are giving AI—and what they are trusting it to produce? Are they even aware their staff are using it? What are the risks?

Unedited responses from Newspage experts below.

5 responses from the Newspage community

Copy all

Star Quote
Copy

Editor
Zac SanterVerified
Director at South Digital
We use AI heavily throughout our agency, and I think it is phenomenal. But there is a difference between employees individually using AI and a business properly adopting it.

Business adoption means knowing which tasks AI is performing, what information it can access, how its output will be verified and which human remains accountable. If everyone is prompting independently and nobody knows what data is leaving the business or how results are being checked, that is not an AI strategy. It is unmanaged outsourcing.

I do not think blanket bans are the answer; they will often drive usage underground. A useful policy should tell people what information must never be shared, when a primary source or conventional tool is more appropriate, what work requires human verification and who owns the final decision.
Star Quote
Copy

Editor
Kate UnderwoodVerified
Founder & Chief People Strategist at Kate Underwood HR and Training
Honestly? I've seen this one up close. A client of mine, lovely little firm, found out one of their team had been dropping the whole staff list into a free AI account. Names, salaries, home addresses, the lot. All to build a rota a bit quicker. No policy, no bad intent, just someone trying to get their Friday afternoon back. And nobody had told them not to, because nobody had thought about it. That's the bit that gets me. Everyone's panicking about AI stealing jobs while it quietly walks in the side door and helps itself to the filing cabinet. Was it useful? Massively. The rota took ten minutes. Trouble is, that data now sits on a server they'll never see again. Ban it outright and people just do it in secret. Pretend it isn't happening and one day it bites you. You don't need a 40 page policy. One page will do. What's fine to put in, what's absolutely not, and who checks it before a customer ever sees it. These are a hot product at the moment.
Copy

Editor
Harvey DhillonVerified
Founder & CEO at Zmartly
A blanket ban on AI doesn't stop people using it. If the tool is on someone's own phone, a ban just moves the use out of sight. For most businesses that makes a ban the wrong call, and it isn't enforceable anyway. If customer information goes into a personal AI account, the business is still responsible for it, and once it's in a public tool, control of it is gone. A practical policy is short: which tools are allowed, staff saying when they've used one, training before they use one, and keeping the prompts so the work can be checked. Responsibility doesn't move onto the software when unchecked AI work reaches a customer. In tax, a careless mistake filed on your behalf can still bring you the penalty, unless you satisfy HMRC you took reasonable care. The business that sent the work out is the one answerable for it.
Copy

Editor
Kelly SmallcombeVerified
Fractional Chief People Officer at Meliorem HR Consultancy
At Meliorem we’re obsessed with AI. We use multiple platforms and let AI do the graft while we do the craft.

The gap isn’t AI adoption, it’s governance. Employees are already using AI daily. Most businesses have no policy, or one that misses how staff actually use it, because nobody asked.

Employees don’t wait for permission when a tool makes their job easier. They find it, use it, and the business often only finds out later, when something looks off. By then, sensitive information may have passed through a personal AI account nobody sanctioned.

Bans don’t fix this, they just remove visibility. Businesses handling it well got ahead with a clear policy, embedding AI into practice and showing staff what good looks like. That’s what keeps unchecked AI out of client work, and keeps the business protected, productive and commercially credible.
Copy

Editor
Paul DenleyVerified
CEO at Oakham Wealth Management
Artificial Intelligence didn't enter firms through procurement - it arrived through people's browsers, and that's where the exposure lies.

The risk isn't someone using AI to draft a note, it's client information being pasted into a free consumer account. A blanket ban is not a reasonable answer given the direction of travel. It may not stop the behaviour, it is more likely to simply push it somewhere less visible.

Far better to approve a small number of enterprise grade tools, be explicit about what can and cannot be entered, and train people properly.

A workable AI policy should answer four questions: which tools are approved, what data is off limits, what level of human review is required, and who signs off before work reaches a customer. "The AI wrote it" is no defence - responsibility remains with the person and organisation relying on the output.