Managers hide their AI use more than their teams do. Ask a team how it uses artificial intelligence (AI) and you will get an answer. How honest it is may depend on whether the person asking admits to their own.
Charter, a US research and media organisation focused on the future of work, surveyed 5,002 full-time US desk-based employees who use AI at work, in a study fielded in August and published on 21 October. Of those, 35.9 per cent had hidden their AI use from a manager, colleague or client at least once. The split by role cuts against the idea that secret AI use is a junior habit. Managers concealed it most, at 43.7 per cent. Executives followed at 37.1 per cent. Workers without management responsibility came last at 27.1 per cent.
In most adoption programmes, managers are also the people asked to report on how their teams use AI.
Why Managers Hide Their AI Use
Among those who concealed their use, 49.3 per cent feared being seen as less capable, 35.5 per cent feared looking lazy or cutting corners, and 26 per cent did not know whether AI was allowed. A manager carrying the first fear has more to lose, because the role rests on judgement.
Silence at that level sets a norm. When the person whose opinion shapes the next review never mentions AI, the topic drops out of normal conversation. A financial analyst in the hospitality industry, one of Charter’s interviewees, described exactly that: “I use it out in the open. We just don’t talk about it,” as Bill Murphy Jr. reported for Inc., the US business magazine.
Falling Resistance Can Be Measured Silence
KPMG’s 2026 Global CEO Outlook, the professional services firm’s annual survey of 1,400 chief executives, was released on 15 October. It placed workforce resistance to AI last of 13 adoption challenges, cited by 26 per cent of chief executives, down from 33 per cent citing employee reluctance or ability to adapt in 2025.
The wording shifted between years, the figures record chief executives’ perceptions, and Charter asked different people different questions, so neither study proves the other. Read together, they raise an awkward possibility: some of that fall in resistance may be people who stopped saying what they think. Objections nobody voices do not appear in anyone’s figures.
The same KPMG study found 56 per cent of chief executives name skills gaps as their biggest workforce challenge. Concealment makes that gap harder to close, because a team that never sees how its manager checks an AI draft is less likely to learn the checking.
What Concealment Breaks
Hidden use cuts the feedback loop in both directions. What works stays with the person who found it, and what goes wrong never reaches anyone who could fix the process. Charter found that only 50.5 per cent of AI users say their organisation has formal AI guidelines. Where guidelines exist, 54 per cent of managers rate them very or completely clear, against 33 per cent of workers. Charter also names concealment as the factor most associated with higher burnout odds.
Another policy document will not close this on its own. Charter ranked AI policies and guidelines as the least useful resource it measured.
Disclosure Starts at the Top of the Room
The fix begins with the people who are hiding most. Leaders publish their own use first: what they use AI for, what they check before trusting it, and where they stopped trusting it. A two-minute example at each weekly team meeting shows what safe disclosure looks like, from the person whose opinion matters.
Then change the measure. Charter describes a demonstration gap: managers are twice as likely to say they are doing what helps their teams as employees are to recall them doing it. Add one question to the next adoption survey, asking whether your manager has shown you how they use AI, and track the gap between that answer and the managers’ own reports. A widening gap warns that the headline numbers are drifting.
In October, Lighting Up The Tunnel (LUTT) argued that AI rollouts are won or lost one layer below the steering committee. The Charter data puts a number on that layer’s silence.
The first honest data point in most adoption programmes will come from a manager saying, in front of the team, what they used AI for this week.
