New research from BambooHR puts a name to something a lot of business leaders have quietly sensed but haven’t wanted to say out loud: the way most companies are adopting AI right now is creating a hidden liability, and workers are the ones carrying it.

BambooHR’s State of the Workforce 2026 report surveyed more than 1,200 employees and business leaders across six industries — construction, technology, education, healthcare, finance, and food and beverage — between March and April 2026. The findings paint a picture of an AI rollout that looks successful on paper but is accumulating what the researchers call “dignity debt.”

What Is Dignity Debt?

The report defines dignity debt as the widening disconnect between rising productivity expectations and the actual employee experience. In plain terms: companies are squeezing more output from their people by adding AI tools, but they’re not redesigning the work, training the people properly, or being transparent about what’s changing. The result is a workforce that’s producing more and feeling worse.

The numbers are striking. While 81% of leaders report a productivity increase from AI adoption, nearly half of them — 49% — also say AI hasn’t actually delivered tangible value and is overhyped. That’s not a contradiction. It means a lot of leaders are seeing short-term output gains while quietly doubting whether any of it adds up to something real.

On the employee side, the picture is bleaker. 85% of workers report daily stress. 29% say they cannot make ends meet despite working full time. And 81% — more than four in five — are considering leaving their careers entirely.

The Skills Gap That Leaders Are Ignoring

One of the sharpest findings in the report is the gap between how leaders perceive their workforce’s AI readiness and how employees actually feel. 74% of leaders believe their employees already have the skills needed for an AI-enabled workplace. But the employee data tells a completely different story — widespread disruption, stress, and uncertainty.

This matters because it means a lot of companies are pressing ahead with AI adoption based on a false assumption: that their people are already prepared. When that assumption is wrong, it doesn’t slow the rollout. It just adds to the pressure employees feel.

The gap is made more serious by the fact that 57% of leaders say they would fire employees who refuse to adopt AI. That’s not a training culture. That’s a compliance culture, and compliance without understanding creates exactly the kind of hidden liability the report is describing.

What Workers Actually Want

The research is clear on what employees want most from their employers through this period: transparency. Nearly nine in ten workers — 89% — said they want greater honesty and visibility into what leadership is planning, what’s changing, and why.

That’s not a complex ask. It’s the baseline that most organizations are failing to meet.

39% of companies have already reduced headcount in the past year because of AI. Workers know this is happening. They’re watching colleagues disappear and absorbing the remaining workload while being told that AI is helping. The trust gap that creates is exactly what dignity debt is built from.

What This Means for Business

The BambooHR findings are a useful reality check for any organization in the middle of an AI rollout. Productivity metrics are going up. That’s real. But if the human side of the equation is deteriorating — if people are stressed, undertrained, financially stretched, and quietly planning their exit — then the productivity gains are borrowed, not built.

Three things jump out from this research as immediately actionable:

Close the skills assumption gap. If your leaders believe the workforce is AI-ready and your employees disagree, that belief is the problem. Proper training — not a one-hour onboarding video, but structured upskilling that builds real capability — is what turns AI tools into productivity that lasts.

Redesign the work, not just the tools. Adding AI on top of existing workflows without changing how those workflows are structured tends to add complexity, not reduce it. The teams seeing genuine gains are the ones that rebuilt the work around what AI can do, not the ones that bolted tools onto what they were already doing.

Be transparent about the direction. Workers are already living through the uncertainty. Pretending it isn’t there doesn’t reduce it. Leaders who explain what’s changing, why, and what it means for people’s roles consistently get better adoption and lower attrition.

For businesses thinking about AI adoption — whether that’s deploying agent workflows, building internal tools, or rolling out AI-assisted processes — the lesson here is that the human layer isn’t optional. Getting it right is what separates sustainable productivity from a short-term spike followed by a talent problem.

If you’re working through what a responsible AI adoption looks like for your business, the Omni Advisory service helps teams think through both the technical and the human side of the transition. And if upskilling your team is part of the plan, Enterprise DNA Learn has structured programs built for exactly this moment.