When employees hide their AI use, disengage, or quietly resist, the problem is rarely the individual. It is a signal the organization has not yet addressed.
Forty percent of employees currently fear that AI will cost them their job. That figure comes from Mercer’s 2026 research, up from twenty-eight percent two years ago. In the same study, sixty-two percent of employees felt their leaders were underestimating the emotional and psychological impact of AI on their teams. Only nineteen percent of HR leaders said they were factoring those impacts into their digital implementation strategy.
These numbers describe a gap that organizations are paying for in ways that do not show up neatly on a dashboard: in disengagement, in resistance, in the fifty-seven percent of employees who report hiding their AI use and passing off AI-generated work as entirely their own. Anxiety about AI is not a personality trait to be managed. It is a workforce signal, and it is worth understanding what it is telling you.
What Anxiety Actually Looks Like at Work
AI anxiety rarely presents as open refusal. In the teams we work with, it tends to look quieter than that, and that is part of what makes it easy to miss.
Some employees become performative users: they adopt AI visibly to signal compliance, without developing any real competence or critical judgment. Others disengage from the parts of their work that involve AI, finding reasons to route around it. A significant number, as the research confirms, simply hide what they are doing, which means neither the employee nor the organization is learning anything useful from the experience.
The common thread is that these employees have not been given a credible answer to the question underneath all the behavior: what does this mean for me, my role, and my value in this organization? Until that question is addressed directly, the anxiety does not go away. It finds expression in the behaviors that quietly undermine the performance outcomes AI was supposed to support.
The Organizational Conditions That Amplify It
Anxiety about AI is not created equally across teams. Research from ResearchGate identifies a consistent pattern: employees who perceive AI as a threat tend toward avoidance behaviors, while those who perceive it as a challenge tend toward engagement. The difference between those two perceptions has very little to do with the individual and a great deal to do with the conditions the organization has built around the transition.
Teams with low psychological safety show higher rates of avoidance. When employees do not feel safe to experiment, ask questions, or flag when AI outputs are wrong, they default to performance rather than learning. The early adopters experiment in private. The skeptics find reasons not to engage. And the gap between those two groups widens over time, creating exactly the kind of fragmented, uneven adoption that makes AI implementation so difficult to assess.
The absence of clear communication about what AI will and will not change also amplifies anxiety. When leaders say AI is important without saying what it means in practice for a specific team or role, the gap is filled by speculation. In environments already under pressure, that speculation tends toward the negative.
Addressing It Without Amplifying It
A common response to AI anxiety is to communicate more: more reassurance, more messaging about opportunity, more emphasis on the benefits. This rarely resolves the underlying concern, and sometimes it compounds it, because it signals that leadership is aware of the anxiety but is responding with messaging rather than with substance.
What reduces anxiety reliably is specificity and involvement. Employees who understand concretely what tasks will change, what support is available, and what the new expectations look like report significantly lower anxiety than those receiving general reassurance. Teams that are involved in defining how AI will be used in their context, rather than having it handed down to them, develop a sense of agency that reframes the transition from something happening to them to something they are navigating together.
This is work that managers need to be equipped to lead. In our experience, the anxiety gap between teams in the same organization often comes down to whether the manager has the tools and the confidence to have these conversations. When they do, the conditions change quickly.
What the Signal Is Asking For
AI anxiety at scale is asking for clarity, honesty, and time. Clarity about what the transition means for specific roles and teams. Honesty from leaders about what they know and what they do not. Time for people to develop competence and confidence at a pace that does not create the cognitive overload that drives avoidance in the first place.
None of that requires waiting for the technology to be fully implemented. The conversations, the preparation, and the habit-building can begin now, and the organizations that start them early consistently find the transition smoother when the full implementation arrives. The signal is worth listening to. It is pointing directly at the work that needs to be done.