AI deployments can deliver massive gains for organizations if they implement the right governance practices, adopt the right procurement policies, and identify the right use cases. The downsides include potentially massive token overruns, wasted implementation costs — and the fact that a significant portion of the workforce will likely attempt to sabotage it.
Yes, “sabotage.”
It turns out that around a third of employees surveyed admit to actively sabotaging their employers’ AI rollouts. A recent article in Futurism reported that:
“[A] whopping 29 percent of workers admitted to sabotaging their company’s AI by entering proprietary info into public AI chatbots, using unapproved AI tools, or intentionally using low-quality AI output in their work without fixing it.”
It’s a safe bet that intentional sabotage isn’t on most tech leaders’ radar when planning AI rollouts. But the reality is that workforces that have negative experiences with or perceptions of AI won’t just avoid using it — they’ll go out of their way to ensure AI fails to deliver.
Tech giants like Google, Meta, and Amazon have made AI use a requirement, and many smaller organizations have followed suit. It’s easy to see why: Not only have these organizations invested heavily in AI solutions, but C-suites believe AI will increase productivity and save costs.
Unfortunately, a significant portion of employees don’t share their optimism.
A global survey by SAP subsidiary WalkMe found that more than half (54%) of employees have bypassed their organizations’ AI tools in the past 30 days and did their work the old-fashioned way instead. Another third of employees haven’t even used AI at all. Worse yet, about 80% of enterprise workers surveyed by WalkMe said that they either avoid using AI when possible or refuse to use it at all.
There are many reasons for this lack of engagement with AI: WalkMe’s survey found that 55% of workers only trust AI with simple tasks, while 45% said AI only gives generic answers. A third said that AI makes work more complicated than doing it themselves. This experience is so widespread, the survey found, that workers report losing a full work day — 7.9 hours to be exact — each week due to friction caused by AI tools.
With these numbers in mind, it’s easy to see why today’s workers aren’t all-in on AI: It makes their jobs harder, it makes routine tasks take longer, and it costs them about 52 working days’ worth of productivity each year.
So, when CEOs issue mandates requiring these workers to use AI on the job, they’re essentially treating a symptom — in this case, AI disengagement or avoidance — rather than the underlying disease.
There’s another reason workers are hesitant to use AI: CEOs, politicians, and a surprising number of commencement speakers have told them that AI is going to take their jobs, while their workplaces simultaneously issue mandates telling them to adopt it themselves. It’s no wonder that EY found that about half of employees are more concerned about AI than they were a year ago.
This anxiety is particularly acute among Gen Z workers and baby boomers. The fact that this anxiety is bookended by the oldest and youngest generations at work today may seem surprising, but it shouldn’t be. After all, both generations are at greater risk than millennials and Gen X workers to have their roles disrupted by AI.
Gen Z, in fact, is so anxious about AI taking their jobs that 44% of them admit to actively sabotaging AI deployments at work.
However, workers aren’t the only professionals experiencing profound anxiety about AI use. Nearly three quarters of executives reported that their organizations’ AI strategies are causing them stress and anxiety. Executives are the ones tasked by boards, investors, and other C-suite leaders with deriving value from these costly AI solutions — a task that becomes impossible when workers avoid or outright sabotage those solutions.
Herein lies the AI anxiety cycle:
Rather than achieving ROI and fostering greater productivity, organizations may instead see lower morale, greater turnover, and lasting damage to their employer brands — and their bottom lines.
Despite the speed at which it’s advancing, AI technology is still very new, and friction and anxiety are bound to occur. But the answer isn't to mandate past the friction — it's to understand and then explain it.
In some industries, as many as 50% of workers suspect that adopting AI is ultimately undercutting their own job security. That suspicion doesn't emerge in a vacuum. It grows in the absence of direct, honest communication from leadership about what AI is being deployed for, who it’s being deployed for, and what it means for the people doing the work.
Fortunately, there’s ample evidence that honest, specific communication changes outcomes. PwC's 2025 Global Workforce Hopes & Fears Survey found that daily GenAI users are significantly more likely to report productivity gains, greater job security, and salary increases than workers who use it infrequently or not at all. Daily users are also far more optimistic about their futures (69%) compared to workers who don't use AI at all (44%). The pathway from anxiety to engagement runs through exposure, and communication is often what opens the door to that exposure.
If communication is half of that equation, psychological safety makes up the other half. An MIT Technology Review analysis found that 83% of executives believe a culture of psychological safety measurably improves the success of AI initiatives, and four in five leaders agree that organizations with that safety in place are more successful at AI adoption overall. PwC advises leaders to give workers the skills and working conditions to thrive alongside AI, help them identify what roles they can fill in the AI age, and explicitly encourage them to "override AI when necessary" without fear of repercussions.
The organizations failing at this aren't doing so because they chose the wrong tools. They're failing because they left a communication vacuum that workers filled with workarounds and quiet resistance.
This doesn't mean that organizations should stop adopting AI. It simply means they need to be clearer and more intentional in how they communicate about it.
Knowing that better communication matters isn't the same as knowing what to say. The leaders getting this right are being specific, honest, and consistent in ways their peers aren't.
Workers are unlikely to respond to messages like "we're becoming an AI-first organization." They’re much more likely to respond to "AI will handle first-draft generation for these three workflows, and here's what that means for your role." Specificity matters because anxiety grows in the space where imagination fills what information doesn't. When workers know exactly what AI is — and isn't — being used for, they're operating on facts rather than worst-case projections.
Acknowledge What You Don't Know
PwC's 2025 Hopes & Fears research found that acknowledging the uncertain future — particularly for entry-level workers — is a key driver of motivation. Most workers aren't expecting leaders to have every answer. They're watching for whether leaders are willing to be honest. Executives who project certainty they don't have about AI's impact on headcount or job scope tend to accelerate the distrust they're trying to avoid.
Build Feedback Loops, Not Just Announcements
Workers experiencing friction with AI tools need somewhere to take it other than quiet resistance. Structured feedback channels — team-level retrospectives, anonymous reporting mechanisms, dedicated AI office hours — keep avoidance from curdling into something harder to reverse. Closing the loop matters just as much: tell workers what changed as a result of their feedback, and when nothing changed, explain why.
Communicate Policy Clearly and Repeatedly
Only about one in four workers says their organization has clear, enforced AI use policies, according to Axios HQ's 2026 analysis — and that shortfall of guidance ranks as one of the top blockers to adoption. That gap doesn't produce neutrality, but shadow AI use, workarounds, and the exact behaviors — entering proprietary data into unapproved tools, using AI output without review — that create security risk and erode the productivity gains AI was meant to deliver.
Make Hands-On Exposure Part of the Strategy
PwC addressed AI anxiety among its own workforce by hosting "prompting parties,” or low-stakes sessions where employees could experiment with AI tools without the pressure of performance evaluations. The approach reflects a principle PwC’s own research supports: Workers who use AI daily are more productive, more optimistic, and less anxious than those who rarely or never do. Communication creates the conditions for that first-hand experience to happen. Without it, workers are left navigating new tools alone — and most won't.
Segment the Message
Gen Z and baby boomers are experiencing this moment differently than millennials and Gen X workers. Gen Z workers fear being displaced before they've built the career equity that gives mid-career workers some insulation, while baby boomers fear that decades of accumulated expertise will be devalued overnight. The same all-hands memo won't serve either group well. Leaders who acknowledge these differences — in targeted communications by role, by team, or by career stage — signal that they understand what their workforce is facing, not just what they wish it were facing.
Takeaways for Tech Leaders
Many of today’s tech leaders are investing in AI deployments, only to have workers avoid or even sabotage those deployments. The damage this can cause goes far beyond a budgetary line item: It can lead to turnover, an inability to innovate, poor productivity, and even damage to brand perception.
To avoid it, leaders must:
Leaders have to walk a tightrope, balancing pressure from above to achieve results with the anxieties of their workforces. It isn’t easy, but a little understanding — and a lot of communication — will go a long way.
To get more expert insights like these on AI, innovation, leadership, and more, visit our resources page today.