There’s a double bind dealing with workers with regards to synthetic intelligence. They’re anticipated to make use of it to be extra productive, but judged on how they accomplish that. Latest Deloitte analysis discovered that almost two-thirds (63 %) of the UK workforce has now used generative AI, a 3rd of these customers accomplish that with out their employer’s information, and one in ten have used an AI software that their employer has banned or would disapprove of. Separate Deloitte figures discovered present that one in six are even paying for AI instruments they use at work out of their very own pocket.
The instinctive response for some employers is to deal with this as a governance downside, with tighter insurance policies, blocked websites and extra monitoring. The Nationwide Cyber Safety Centre (NCSC) has additionally warned, with good cause, in regards to the dangers of shadow AI going unchecked.
But framing this purely as rule-breaking misreads what’s occurring. It doesn’t deal with why workers really feel they should conceal AI use within the first place, and it dangers making issues worse.
The psychological toll of concealment
A typical assumption is that workers conceal their AI use as a result of they’re chopping corners or passing off AI-assisted work as their very own. Whereas there’s some reality in that, it’s removed from the entire image.
A lot of the urge to hide is rooted in nervousness. There’s the worry of being changed, with two-thirds of Deloitte’s weekly AI customers fearful their managers will assume AI can do their job. There’s the worry of being judged, with almost 1 / 4 citing the stigma hooked up to utilizing it. And there’s the sense of being left to work it out alone – solely round a 3rd say their leaders discuss AI with an excellent understanding, and half have had no coaching.
These fears aren’t unfounded. In a single US experiment, Atlassian discovered that folks rated a colleague who disclosed their AI use as ten occasions lazier than one who didn’t, and have been far much less more likely to suggest them for a high-visibility mission – even when the work was similar.
The nervousness doesn’t ease as individuals get extra out of AI, both. Anthropic’s analysis discovered that folks within the roles most uncovered to AI are thrice extra more likely to fear about being displaced by it than these within the least uncovered roles, and people reporting the most important productiveness features from AI are additionally among the many most involved.
It additionally falls hardest on youthful employees, who’re essentially the most fearful about displacement. These are the individuals who most want steering, as a result of they’re nonetheless constructing the abilities and judgement their careers will rely on. In the event that they don’t really feel protected speaking about how they use AI, they’re much less more likely to get that steering, and fewer more likely to be taught when to depend on AI and when to not.
The value of secrecy
Heavy-handed bans and surveillance are likely to feed these fears reasonably than resolve them, which solely pushes AI use additional out of sight. As one evaluation of shadow AI put it: “Bans not often get rid of utilization. They merely drive it underground.”
As soon as AI use goes underground, organisations lose the openness they most want throughout a interval of speedy change. Data stops spreading, as a result of no person shares what works. Errors go unchallenged, as a result of no person admits how a chunk of labor was produced. And the individuals designing roles, coaching, methods and workspaces lose sight of how work is definitely being accomplished.
Even the NCSC isn’t telling individuals to dam each software. It advises organisations to undertake “a constructive cyber safety tradition” with an open dialogue. Deloitte reached an identical conclusion, arguing that the reply isn’t merely stricter guidelines. When a nationwide safety physique and a worldwide agency each say openness beats management, HR leaders have a transparent mandate to behave.
What employers can do
Many individuals use AI freely of their private lives, but maintain again or conceal it at work. This hole tells us that the problem isn’t the expertise itself, however the best way it’s embedded into the office. Round 35 situations must be met to ensure that office AI adoption to achieve success, and half of these are cultural – so clearly this isn’t merely a technological downside. If we are able to repair the tradition, we are able to create the foundations for good-quality, open AI use.
That begins on the prime. Nothing reduces stigma quicker than senior individuals speaking brazenly about how they use AI, together with the place it helped and the place it obtained issues improper. If leaders maintain their very own use hidden, workers will draw the apparent conclusion.
It additionally means wanting past ‘productiveness’ and treating AI as a human improvement concern. Staff need assistance understanding how their roles are altering, the place their judgment issues and which human qualities are price defending. Used effectively, AI clears away repetitive duties to create extra room for creativity and big-picture considering. Used badly, it replaces the very practices via which individuals construct their human expertise and experience.
None of this implies employers have to surrender management. Those who wish to handle synthetic intelligence extra intently can, however guidelines solely work if individuals first really feel psychologically protected, engaged and sorted. With out these situations, tighter management will merely push AI use additional underground. It’s as much as us as employers to place the readability, dialogue and assist in place to make sure that nobody in our organisation feels they’ve to decide on between being productive and being trusted.
Gethin Nadin is chief advisory officer at Benifex and writer of Coronary heart within the Age of the Mechanical. Gethin is an award-winning psychologist, bestselling HR writer, and in 2026 was named one of many UK’s High 100 Most Influential Folks. Gethin has been awarded greater than ten awards since 2024 for his work in influencing the worldwide HR trade, most lately awarded double Gold on the prestigious Stevie’s awards in New York in recognition of a profession devoted to human sources. He was additionally awarded a lifetime achievement award by the UK Worker Expertise Awards in 2025 and named PwC’s ‘Excellent Contribution to the Business 2024’. Gethin is a Fellow on the historic King’s Faculty London, King’s Enterprise College and Chair of the Westminster assume tank, The Coverage Liaison Group on Office Wellbeing.
