A recent study from the Technical University (TU) of Darmstadt upends a comforting assumption long held by labour market experts: that fear of AI-driven job loss stems primarily from a lack of understanding.
Conventional wisdom held that ignorance bred panic. The Darmstadt study shows the exact opposite. The better people understand how AI works, the greater their anxiety about management pressure and job displacement. Around 43% of respondents with a very good understanding of AI expect the technology to soon take over their work. Tech professionals and those closest to digital automation—such as lawyers, software developers, and social media experts—recognise its disruptive power far better than workers in fields less susceptible to immediate automation, such as tradespeople and healthcare workers.
This anxiety is far from unwarranted. According to a Mercer survey conducted in May 2026, 99% of CEOs in Germany expect to lay off workers over the next two years as AI tools take over their tasks.
The Training Gap and Capitalist Contradictions
While AI has rapidly integrated into everyday life, its use remains uneven. Regular usage is heavily concentrated among younger, well-educated, and predominantly male users.
Furthermore, researchers point to a stark “AI training gap”. Just 15% of the 2,000 workers surveyed had received appropriate AI training from their employers, leaving a whopping 85% with zero training. This exposes a classic corporate contradiction: management complains about skill gaps while refusing to invest in workforce training, preferring to offload the cost onto individuals—a hallmark of neoliberal labour markets.
Compounding these economic fears are two major operational risks:
Hallucinations: AI outputs can look entirely convincing while remaining fundamentally incorrect.
Data Privacy: Unanswered questions remain regarding what happens to the data workers produce or the data management collects about workers.
Who Uses AI, and How?
Statistical breakdowns across demographics reveal wide disparities in weekly AI usage:
Generation Z (1997 to 2012): 73.7%
Millennials (1981 to 1996): 62.5%
Generation X (1965 to 1980): 38.8%
Baby Boomers (1946 to 1964): 26.3%
Notably, the sharpest drop in usage occurs in the middle of working life, between Millennials and Generation X (a 24-percentage-point decline). This indicates that early exposure—via schools, universities, or modern vocational training—plays a massive role in whether AI becomes a habitual tool. Gender gaps are narrower (49.3% for men vs. 43.9% for women), while education level plays a steady upward role (ranging from 40% among lower-education groups to 55.2% for higher education).
When people do use AI, their habits vary drastically based on familiarity:
Information & Learning: 71.0% use it for enquiries and fact-checking.
Text & Writing: 51.9% use it to create, revise, or translate texts (spiking to 64.8% among office and knowledge workers).
Brainstorming & Media: 38.1% use it for content generation.
Programming & Data Analysis: Just 16.0% use it for complex technical tasks.
Proximity to AI dictates its application. Workers familiar with the technology use it for demanding tasks, while those more distant limit themselves to basic explanations.
Moving Toward Democratic Control
Ultimately, workers’ concerns are not focused on speculative sci-fi scenarios of superintelligence but on concrete workplace realities: algorithm-driven management, job security, and data protection.
In-depth knowledge does not breed complacency; it sharpens the understanding of coming automation. Because corporate management often finds fearful workers easier to manipulate, business leaders, politicians, and trade unions must address these anxieties head-on. Restoring trust requires more than corporate rhetoric—it demands reliable applications, transparency, responsible data handling, and genuine worker participation.
