Workers Reveal Managerial Eagerness, Unusual Slip-ups and Impending Risk as They Teach AI Their Jobs

The rapid integration of artificial intelligence into workplaces is not merely a technological shift but a profound transformation of the labour market, one that workers on the frontline describe as leaving them feeling “devalued” and warning of a downward trajectory in work quality. This sentiment emerges as global economic bodies sound the alarm, with International Monetary Fund managing director Kristalina Georgieva stating the phenomenon is “like a tsunami hitting the labour market.”

The Scale of Disruption

Recent IMF analysis confirms the wave’s size, estimating that nearly 40% of global employment is exposed to AI-driven change, a figure that rises to 60% in advanced economies due to their concentration of cognitive roles. While the technology holds the potential to accelerate global growth by 0.8%, its impact on jobs is deeply dual-sided. The IMF notes that roughly half of impacted jobs may see productivity enhanced by AI integration, but for the other half, AI applications could execute key tasks currently performed by humans. A World Economic Forum report projects that by 2025, AI will have displaced 75 million jobs globally but created 133 million new ones, though the transition promises significant disruption for many.

The effects are already unevenly distributed. Analysis indicates a disproportionate impact on early-career employees. For instance, employment for workers aged 22-25 in highly AI-exposed sectors fell by 6% between late 2022 and July 2025, while employment for older workers in those same sectors grew. This disparity is linked to the nature of AI’s capabilities; it tends to automate codified knowledge—the “book learning” common among recent graduates—but not the tacit knowledge gained through years of experience. Consequently, AI may substitute for entry-level roles while augmenting the work of seasoned professionals, a dynamic reflected in wage trends where occupations with a high “experience premium” show increased wage growth alongside AI exposure.

Training the Replacement

For many workers, this abstract economic force has a very personal cost. Christie, a 55-year-old academic editor based in the UK, found herself unwittingly training the AI system that would later undermine her livelihood. She was asked to help train new “assistant editors,” only to discover via a company newsletter months later that they were an AI programme. “I now earn less money for correcting the mistakes of an AI, which takes me longer than editing from scratch,” she says, describing feeling “devalued, betrayed, and furious.” Her experience points to a broader trend in academic publishing, where AI is used for pre-editing to reduce costs, though the output often requires heavy human review to avoid factual errors and misrepresentation.

This pattern of workers being tasked with refining their own potential replacements is repeated elsewhere. Philip, a 45-year-old translator in New Jersey, has spent four years correcting AI-based translation engines his supervisors aim to adopt. “Even after years of this, besides tending to produce formulaic results, they are still unreliable and inadequately accurate,” he says, arguing the overall effect is a decline in quality. Similarly, Joe, a 50-year-old award-winning marketing writer from Milwaukee, was laid off in August 2025, just two weeks after handing in extensive “AI process workflows” he had been asked to develop. “Working for this company and being asked to do this – training your robot replacement – feels like digging your own digital grave,” he says, noting his former workload is now delegated to junior employees following his AI documentation.

Implementation Hurdles and Human Nuance

The drive to implement AI, which some workers perceive as a form of “groupthink” within companies, often runs aground on the complexities of human communication and data quality. In healthcare, Professor Mark Taubert, a palliative care consultant at Velindre University NHS Trust in Cardiff, collaborated on a pilot AI chatbot named Rita designed to answer patient questions out of hours. While about 50% of its responses were accurate, it struggled with the vagaries of human pronunciation, misspellings, dialects, and jargon—such as patients saying “morphium” instead of morphine. The project, used “with a lot of caveats and warnings,” was eventually discontinued after funding ended and following recommendations from an external evaluation. Another AI tool named Rita is being developed in Hong Kong to train healthcare workers, highlighting the sector’s cautious exploration of the technology.

These struggles underscore a critical barrier for AI initiatives: poor data quality. Gartner predicts that 60% of AI projects will be abandoned by 2026 due to a lack of AI-ready data, with inaccurate or biased information leading to potential financial losses and discriminatory outcomes. Furthermore, many AI systems operate as “black boxes,” lacking transparency and explainability, which complicates compliance and erodes trust. Ethical concerns extend to privacy violations, surveillance, and the risk of AI increasing dishonest behavior.

A Changed Future of Work

Looking ahead, the very nature of expert roles is poised for change. Filippo, a 44-year-old associate professor of mathematics in France collaborating with startups on AI that can prove theorems, observes that while results are still limited, the tools are strengthening daily. “With most of my colleagues experimenting with this AI technology, we are convinced that a mathematician’s work will look completely different in 10 years’ time, or perhaps even less,” he says. AI is expected to handle mundane tasks, freeing humans for more complex work—a potential augmentation echoed in fields like nursing, where AI could take over rote duties to allow more patient care.

However, for workers like Christie, trapped in a “toxic cycle” of correcting AI errors to pay rent, or Joe, considering a career pivot into sales amid fears of another layoff, the immediate reality is one of uncertainty and devaluation. Their experiences, set against a backdrop of global economic recalibration, illustrate that the promise of AI-driven productivity is, for many, tempered by a profound and personal disruption to their working lives.