Insights on the world of work | Insights on the world of work

AI Is Everywhere at Work—Confidence Isn’t

Something strange is happening in the AI adoption numbers: usage is up, but people feel worse about it.

Regular AI usage among workers jumped 13% over the past year according to ManpowerGroup's 2026 Global Talent Barometer. In the same time frame, confidence in using those tools dropped 18%. The fact is employees are clicking the buttons and submitting the prompts. But three-quarters of them don't feel like they're doing it well—or even feel comfortable doing it.

Something is off in how AI is being rolled out.

The data shows a clear pattern: as adoption rises, so does anxiety about job automation, creating what looks like an AI adoption paradox. And it's showing up in boardrooms and team meetings everywhere.

Seema Daryanani People and culture partner at Gemini App, Google DeepMind

Seema Daryanani of Google DeepMind, put it plainly at a recent panel on AI on the future of work with From Day One: “On one hand, there's a lot of excitement around transformation, but I also see a lot of fear. It's this interesting balance of excitement and fear that I see.”

That balance isn't unique to any organization. It's the defining tension of this moment.

Understanding what sits beneath it matters—because the organizations that get this right aren't just deploying AI faster. They're building workplaces that are more resilient, more adaptive, and more human.

The Human Element

The resistance isn't a skill gap or about the tech. It's a human problem.

Many organizations frame AI adoption as a training problem. The theory: if people understood the tools better, used them more, and got past their discomfort, adoption would follow.

But MIT research published in 2025 suggests this framing is why 95% of corporate AI initiatives fail to meet their objectives. When AI arrives in a team's workflow, the surface-level message is about efficiency and speed. The subtext—heard clearly by many workers—is that the expertise they've spent years building may no longer mean what it once did. That's not irrational. It's a gut response to a rapidly shifting environment.

The fear hits experienced workers hardest:

Baby Boomers saw a 35% decrease in AI confidence over the past year
Gen X dropped 25%

These aren't people resisting change out of stubbornness. They're professionals who built hard-won expertise over decades, now watching it become potentially irrelevant overnight, with no clear path to rebuild.

Meanwhile, a February 2026 Gallup survey of nearly 24,000 U.S. employees found: 27% of workers in AI-adopting organizations report disruptive workplace change compared to 17% in organizations that haven't adopted AI

Disruption is the lived experience even when the intent is transformation.

From Hesitation to Adoption

The research points consistently to one underlying factor: psychological safety. It’s not about training completion rates, tool access, or leadership mandates.The real question employees are asking is quieter and more fundamental: Is it safe to not know what I'm doing yet?

According to MIT Technology Review research, psychological barriers—not technological ones—are the greater obstacle to enterprise AI adoption. Nearly a quarter of employees hesitate to lead AI projects for fear of being blamed if something goes wrong. Organizations that simply deploy tools see low adoption. Companies that redesign performance metrics, reward experimentation, and encourage visible leadership engagement see both literacy and adoption rise

The difference isn't the technology. It's the culture built around it.

When Leaders Lead

The leaders helping their organizations move through fear aren't doing it with better slide decks.

They're going first. They use the tools visibly. They share what worked—and what didn’t. They create conditions where others feel safe to do the same.

Daryanani is one example. During the panel, she shared: “I started building an agent yesterday. I chose to build a chief of staff. And I can't wait until it's done.” She wasn't announcing a corporate initiative. She was describing something personal—something practical—something experimental. That kind of self-directed, hands-on engagement is exactly what research identifies as the strongest driver of real adoption.

Closing the Gap

The fear around AI is real. The disruption is real. But momentum alone won’t close the gap between use and confidence. That gap narrows when organizations stop treating AI as a rollout, start treating it as a transition and give people room to learn, experiment, and not have all the answers

That’s when adoption shifts from obligation to ownership. AI adoption isn’t just about tools. It’s about people.