AI is being deployed across organisations on an ever-larger scale. The technology is expected to support employees, accelerate workflows and improve productivity. However, recent research by digital transformation consultancy Adaptavist suggests that many knowledge workers are experiencing a different reality.
The study surveyed 2,500 knowledge workers across the United Kingdom, the United States, Canada, Germany and Spain. The findings point to growing AI fatigue, increasing concerns about the quality of AI-generated output and a rising amount of verification work. Employees describe what the report refers to as the AI verification tax: the additional time and effort required to review, validate and correct AI-generated output.
Longing for the pre-AI workplace
Nearly two-thirds of respondents (65%) say they regularly miss the way work was done before AI became widely adopted. More than one-third (38%) would remove AI from their workplace altogether if given the opportunity. These findings do not suggest that employees reject AI outright. Rather, they indicate that many workers do not experience its introduction as an unequivocal improvement to their daily work.
Work Is becoming less meaningful
According to respondents, the quality of AI-generated output is also affecting how they experience their work. Almost half (46%) report that dealing with low-quality AI output has made their work more repetitive, while 37% say AI has reduced their sense of engagement with their job. AI can generate large volumes of text, information and other content within seconds. However, when that output lacks reliability or relevance, employees must invest significant time reviewing, correcting and refining it. As a result, work increasingly shifts from creating original content to evaluating and repairing AI-generated material.
The AI verification tax
One of the study’s most significant findings is that AI does not always deliver the anticipated productivity gains. Forty-two percent of respondents report spending more time verifying AI-generated output than the amount of time AI actually saves them. Almost half (49%) say poor-quality AI output actively delays projects, while 55% believe it reduces overall team efficiency. These findings illustrate that generating content alone does not automatically create productivity gains. If employees must devote substantial time to identifying inaccuracies, correcting mistakes and improving incomplete output, the overall workflow may ultimately become slower rather than faster.
Concerns about creativity, misuse and privacy
Among employees who would prefer to eliminate AI from the workplace, several concerns stand out. Thirty-one percent believe AI reduces creativity. Another 29% are concerned about potential misuse of the technology, while 28% cite issues relating to surveillance and privacy. These concerns extend beyond the quality of AI-generated output. Employees also worry about the broader implications for creativity, autonomy, responsible AI use and personal privacy.
Employees feel pressure to compete with AI
The research also highlights growing performance pressure. Half of all respondents (50%) feel that their work is being compared directly with AI-generated output, and many do not consider this comparison fair. One in four employees (25%) primarily use AI simply to cope with their existing workload, while 23% rely on it to keep pace with colleagues. AI is therefore no longer used solely as a productivity tool. For many employees, it has become a necessity in order to meet organisational expectations and increasing workplace demands.
Uncertainty about the purpose of AI
A significant challenge identified by the study is that many employees do not fully understand why they are expected to use AI in their role. More than one-third of respondents (36%) say they frequently do not understand the rationale behind mandatory AI use. The same proportion also report experiencing AI fatigue as a consequence. These findings suggest that many organisations may be placing too much emphasis on AI adoption itself and too little on the business problems the technology is intended to solve. AI risks becoming an objective rather than a means of improving work.
Measuring AI usage is not the same as measuring results
According to Neal Riley, AI Innovation Lead at The Adaptavist Group, organisations frequently evaluate AI initiatives primarily through adoption metrics. They measure who is using AI and how often employees interact with the technology. However, these figures reveal little about the actual impact on work. High adoption rates do not necessarily translate into better processes, faster project delivery or higher productivity. Instead of focusing solely on AI usage, organisations should evaluate how AI influences the nature, quality and outcomes of work. This requires a clear understanding of business activities, workflows and value streams. Only then can organisations determine where AI genuinely creates value and where it merely generates additional verification work.
Employees remain positive about AI
Despite these frustrations, employee support for AI remains remarkably strong. More than two-thirds of respondents (67%) would like their organisation to expand its use of AI. In addition, 69% trust that AI is being used ethically within their organisation, while 66% believe their employer has been transparent about its implementation. The findings therefore reveal a clear paradox. Employees are generally supportive of broader AI adoption while remaining highly critical of its practical consequences and the quality of the output it produces.
Careful AI implementation is essential
The research does not conclude that AI is inherently harmful to employees or organisations. Instead, it demonstrates that the way AI is implemented largely determines its impact. When introduced thoughtfully, supported by clear governance and accompanied by appropriate employee guidance, AI has the potential to strengthen work. However, when organisations focus primarily on adoption rates and usage frequency, AI may instead create additional work, quality issues and a less meaningful work experience. The key question, therefore, is not simply how many employees are using AI. Organisations must determine whether AI is genuinely improving the quality, effectiveness and outcomes of work.
Reageer op dit bericht