As CEO of ComPsych — the world’s largest provider of employee mental health services — I spend a lot of time thinking about what’s making workers anxious. Right now, AI is at the top of that list.
According to a 2025 Pew Research Center report, workers are more worried than hopeful about AI in the workplace. And the consequences of that anxiety go well beyond morale: research shows that when employees believe they are likely to lose their job, the risk of serious psychological distress rises considerably — along with the likelihood of mental health leave, disengagement, and burnout. Leaders are rolling out AI tools at speed. Most are treating this as a technology challenge. It isn’t. It’s a people challenge — and getting it wrong has a measurable cost.
The anxiety your employees aren’t telling you about
Job insecurity driven by AI fear manifests in two damaging patterns. The first is disengagement — distracted, uncollaborative workers doing the minimum to get by. The second is its opposite: workers who over-compensate by becoming hypersensitive and over-extended, eventually burning out. Both patterns hurt team performance. Both are symptoms of the same underlying failure: leaders who haven’t addressed what their employees are actually afraid of.
To counteract this, leaders need to focus on building genuine trust. Trust doesn’t just happen — it’s earned. The most effective way to build it is through early, open, and consistent communication. The impulse to wait until everything is figured out before communicating is understandable. But in the absence of communication, rumors form, narratives harden, and fear sets in. Sharing relevant information as soon as it’s available — being transparent about what’s still being determined and when more details will follow — allows leaders to inform and reassure simultaneously.
In the case of AI, this can be as simple as defining a clear company philosophy. Articulating that AI may change parts of a person’s job but does not diminish their value gives leaders the opportunity to set the tone, reduce fear, and support emotional well-being before anxiety takes hold.
Where AI should — and shouldn’t — have the final word
Clarity about where AI should — and shouldn’t — have the final word is one of the most important things a leader can establish early.
I remind our teams constantly: AI has to earn our trust. If you were collaborating with a new colleague for the first time — one with only a few years of real-world experience — you would never take their work, assume it was fully correct, and move forward without carefully reviewing it, providing feedback, or challenging the portions you disagreed with. The same scrutiny applies to anything produced with AI assistance.
Beyond quality checks, there are categories of work where human judgment, ingenuity, and creativity must remain the final word. Every organization will need to define its own guardrails. As a company providing mental health services to millions of people worldwide, we’ve been explicit with our teams: the human expertise of our clinicians is paramount. We are embracing AI to enhance operations and aid patient navigation — but we will never defer to a large language model when someone is in crisis and needs human-centered care.
The overreliance trap — and how to avoid it
Setting clear expectations is necessary but not sufficient — organizations must also actively ensure their employees are growing alongside the technology, not being hollowed out by it. Research from MIT’s Media Lab found that heavy reliance on AI tools can atrophy the independent thinking and problem-solving skills people need most.
It is incumbent on leaders to ensure teams are using these tools to vault their creative and strategic thinking to new heights — not as crutches they eventually won’t be able to function without. This means actively encouraging imaginative, out-of-the-box thinking, reinforcing the value of individual skills, and investing in continuous learning and development programs. Upskilling is not optional: the skills needed to thrive alongside AI, and the tasks that make up our workdays, will change enormously.
The leaders who get this right won’t just be the ones who deploy AI fastest. They’ll be the ones who brought their people along — reducing fear, building trust, and preserving the human judgment that no model can replicate. The goal isn’t to help your workforce adapt to the future of work. It’s to help them build it.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
