80% of workers quietly reject AI as 40% of technology investments go wasted: The truth behind the AI hype

About 80% of workers refuse to use AI even as companies invest tens of millions of USD in the technology each year, according to reporting that highlights a growing disconnect between corporate AI spending and employee adoption.

From “shadow AI” to open avoidance

Earlier, “shadow AI”—employees using tools such as ChatGPT or Claude outside corporate systems—was often viewed as a sign of productivity gains. Staff reportedly used personal accounts to complete tasks faster, cutting work that took hours down to minutes.

However, the dynamic has shifted. A MIT study cited in the report found that more than 90% of employees used personal chatbots in daily work routines, often without approval, while only 40% of companies formally registered AI models. That created an informal “underground economy,” with management viewing it as a risk and employees seeing it as a practical way to get work done more efficiently.

Survey data shows low usage and declining trust

A global survey of 3,750 leaders and employees across 14 countries, conducted by SAP via WalkMe, found that 54% of employees ignored their company’s AI tool in the last 30 days and returned to manual work, while 33% said they have never used AI.

Overall, the report characterizes roughly 80% of the labor force as “avoiding” or refusing AI in their work.

Spending rises, adoption lags

Companies continue to invest heavily. The average digital transformation budget rose 38% to $54.2 million, but 40% of that investment did not deliver results due to failure to adopt AI in practice.

The report also points to a widening perception gap: 61% of leaders believe AI can handle important decisions, but only 9% of employees agree. Up to 88% of leaders say the tools are sufficient, compared with just 21% of employees. The resulting trust gap is described as 52 points, alongside a 67-point tooling gap.

Why employees disengage: effectiveness without the “how”

Economist Steve Hanke of Johns Hopkins argues that AI has not delivered the breakthrough expected, particularly for productivity. He contends that if AI were truly effective, labor productivity would rise markedly.

WalkMe CEO Dan Adika similarly notes that the share of employees who use AI for meaningful work remains below 10%.

Experts use analogies to describe the problem: AI is likened to a Ferrari or a Formula 1 car—powerful, but unable to deliver value without the ability to operate it.

According to Adika, employees lack the “driving” skills (prompting), the “fuel” (data and context), and the “roads” (APIs or proper infrastructure). When these elements are missing, effectiveness is described as nearly zero. Brad Brown of KPMG makes a similar point, saying that without a skilled driver, advanced technology cannot deliver value.

Measurable productivity losses

The report frames the consequences of the gap as increasingly measurable. Employees lose 51 days of work per year due to technology friction—up 42% from 2025—equivalent to 7.9 hours wasted per week.

It also highlights a productivity paradox: Goldman Sachs suggests AI can help users save 40–60 minutes per day, but outcomes vary depending on how people use the tools.

“Quiet quitting” of AI and concerns about accuracy

The report describes a new trend under “shadow AI”: not subverting rules, but choosing not to use AI at all. Adika says employees resist due to professional pride, believing AI is not good enough to replace humans and can always be wrong.

It also cites limitations such as hallucinations, incorrect responses, and inaccuracies as factors that erode trust.

In parallel, some workforce reductions—such as moves reported at Oracle or Block—are met with suspicion by many as “AI-washing,” where AI is used as a pretext to mask other business decisions.

What companies can do: access, training, and role-based support

Experts interviewed in the report argue the core issue is not the technology itself, but how it is approached and used. KPMG categorizes employees into three groups: Builders, Makers, and Power users, with the goal of elevating all staff to higher levels of AI usage.

Notably, one-third of corporate employees have never used AI. The report says this group has the highest anxiety and the least training and support, and that they are not necessarily anti-AI—they simply have not been given proper access.

Bottom line: value depends on combining tools with trust and process

Even skeptics such as Steve Hanke acknowledge that AI can save substantial time when used properly. The report concludes that AI is not a magic fix: it becomes productive only when paired with skills, processes, and trust.

In the competition for adoption and results, the report argues that organizations that effectively combine humans and technology—not necessarily those that invest the most—will be better positioned, while those that fail to close the gap may lose productivity and competitive advantage.