Workers across the United States say they feel caught between adopting productivity tools and training their own replacements, a worry underscored by a recent poll that found roughly 30% of Americans believe their jobs may be made obsolete by artificial intelligence. The anxiety is showing up in career decisions as well: more college students are reportedly switching majors amid concerns about what AI will mean for future employment.
The concern was voiced repeatedly to career coach Erin McGoff, founder of the AdviceWithErin career-education platform. “I have people who say that they, every day, feel like they're training their replacement,” McGoff said, recounting conversations with professionals across sectors. She said the fear is especially acute in roles where workers routinely feed data or correct outputs from generative tools, feeding the perception that the systems are learning directly from them.
Corporate investment in AI helps explain the unease. Companies are pouring billions into the technology to boost efficiency, and some executives have publicly linked headcount reductions to AI initiatives. Analysts warn that these claims sometimes mask other business pressures. JP Gownder, vice president and principal analyst at Forrester Research, accused some companies of “AI-washing” — using AI as a convenient rationale for layoffs driven by post-pandemic over-hiring, higher interest rates and broader economic uncertainty.
At the same time, some firms say they are deploying “digital employees” to take on routine tasks so human staff can focus on higher-value work. Financial firms such as BNY — which has rolled out automated agents to handle mundane processes — have framed the technology as a way to free staff for more complex responsibilities rather than to replace them outright. Yet that message has not entirely calmed workers’ fears that today’s helpers could become tomorrow’s replacements.
Experts argue a wholesale one-for-one substitution remains unlikely in the near term because most jobs require juggling multiple tasks, navigating ambiguity and exercising judgment — capabilities where AI still lags. Gownder pointed out that many deployed AI tools are not designed to learn by passively watching human workflows; they require structured training, labelled datasets and oversight. “We're not close to that, almost in any sphere,” he said, stressing that job change is more probable than mass erasure.
Still, the rising anxiety is having effects on career planning and employer-employee dynamics. Some students are altering degree plans in anticipation of different labour market demands, and employees are increasingly weighing the risks of integrating new tools into their daily work. The trend also raises questions for policymakers and companies about reskilling and transparency: if AI investments are intended to reallocate tasks rather than eliminate roles, firms may need to pair automation with clear retraining pathways.
The debate over AI's labour impact is playing out even as major technology employers pivot resources toward AI development. Recent moves in the sector — including large-scale workforce restructurings tied to AI strategies — have heightened scrutiny of corporate motives and the practical limits of current systems. For now, observers say the most immediate likely outcome is widespread role transformation, not wholesale replacement, leaving workers and institutions to navigate an uncertain transition.