AI is saving workers more than two hours a day. That sounds like an unqualified win, and in many ways, it is. But beneath the productivity headlines, something more complicated is happening. Employees are getting faster, but some are also getting less confident, less skilled, and less certain they can do their jobs without a machine doing much of the thinking for them. That tension is the defining workforce challenge of 2026, and most companies aren’t prepared to address it.
New research from GoTo, conducted in partnership with Workplace Intelligence, surveyed 2,500 global employees and IT leaders on AI use and sentiment. The findings tell a story about a workforce caught between the tools that help them and the habits those tools are forming. Fifty percent of employees now say they rely on AI too much. Thirty percent say they can no longer function without it. And 39% believe their overreliance on AI is actively eroding their skills and making them less intelligent, a number that climbs to 46% among Gen Z workers. These aren’t fringe opinions. They are the quiet consensus of a workforce that adopted AI fast and is now reckoning with the consequences.
The Pressure to Use AI Is Outrunning the Guardrails to Use It Well
One of the clearest findings in the research is how much external pressure is shaping AI behavior at work. Sixty percent of employees say they feel pressured to use AI tools to boost productivity regardless of whether the task calls for it. That pressure, absent the right training and policies, is a setup for misuse.
The numbers bear this out. Seventy percent of employees (up from 54% just a year ago) admit they’ve used AI for sensitive or high-stakes tasks, including legal or compliance work, decisions requiring emotional intelligence, and actions involving confidential information. These are exactly the domains where human judgment is most irreplaceable, and where AI errors carry the highest cost. The fact that this number jumped 16 percentage points in a single year suggests the problem isn’t slowing down on its own.
Compounding this is an “AI workslop” problem that’s starting to tax the entire workforce. Forty-three percent of employees say they’ve submitted AI-generated content despite suspecting it was low quality or contained errors. With that in mind, it’s unsurprising that 77% percent say reviewing AI-generated work takes more time than reviewing human work. And 66%sixty-six percent say wading through other people’s AI output creates extra work for them. The efficiency gains from AI are real, but they’re being partially offset by a flood of under-reviewed, unreliable output that everyone else must spend time, energy, and resources to clean up.
The Leadership Gap Is Where the Real Risk Lives
What makes these findings particularly striking is the disconnect between employees and the leaders responsible for guiding them. Eighty-four percent of employees say their company could do more to encourage responsible AI use, however only 48% of IT leaders agree. That gap of 36 points is a signal that IT leadership is significantly underestimating the extent of the problem.
The policy picture is just as concerning. Only 44% of IT leaders say their company has an AI policy in place at all. And among those that do, 77% of employees say the policy needs improvement. Meanwhile, 80% of employees and 60% of IT leaders acknowledge that most workers aren’t being properly trained to use AI tools. The infrastructure for responsible AI use, including the policies, the training, and the role-specific guidance hasn’t kept pace with how fast employees have adopted these tools.
This is not a technology issue, not a generational issue, and not something that will self-correct as AI matures. Employees are not misusing AI out of laziness or bad faith; they’re doing it because they’ve been handed powerful tools without the context and enablement to use them well, and told implicitly or explicitly to produce results. When organizations reward output without asking how it was produced, they get exactly what they incentivize.
What Companies That Get This Right Will Do Differently
The same research that surfaces these problems also points toward solutions, and they’re not complicated. They require organizational commitment, not technological breakthroughs.
The priority is building AI policies that work. That means policies employees understand, see as relevant to their daily work, and feel equipped to follow, not compliance documents that live on an intranet page. Given that 65% of employees say their employers have not equipped them with the skills they need as AI takes over more work, this must be paired with genuine training investment, including role-specific guidance on where AI adds value and where it doesn’t belong.
The second priority is deliberate investment in human skills. Workers themselves identified the capabilities they believe will matter most in an AI-driven workplace: creative thinking, emotional intelligence, sound judgment, and the ability to know when to trust AI outputs and when to override them. These aren’t soft skills in the dismissive sense; they are the hard-to-automate competencies that determine whether AI amplifies a workforce or quietly hollows it out. They’re also the foundation of effective human-AI collaboration.
The employees who will create the most value aren’t those who use AI the most, but the ones who know how to work alongside it. Workers should focus on contributing the judgment, context, and creativity that AI cannot supply, while letting AI handle the volume, speed, and synthesis it does well. Companies that train employees to operate in that partnership model, rather than simply handing them tools and expecting results, will be better positioned when the next wave of AI capabilities arrives.
The third is cultural: leaders need to model what responsible AI use looks like, not just mandate it. Employees who see their managers using AI thoughtfully, knowing when to rely on it, when to push back on its outputs, and when to set it aside entirely are more likely to develop the same instincts. Policy shapes behavior at the edges; culture shapes it at the center. Eighty-eight percent of employees say AI has benefited them. That number should give every business leader confidence that the technology is working. But the same research makes clear that productivity gains alone are not a strategy. The companies that will win the next decade of work aren’t the ones who pushed AI adoption hardest. They’re the ones who built the organizational discipline to use it wisely, and kept their people capable, confident, and trusted in the process.