The fear of AI replacing jobs is real, it's just not the fear you think

The World Economic Forum's Future of Jobs Report 2025, drawing on data from over 1,000 companies across 55 economies, projects 92 million jobs displaced globally by 2030, offset by 170 million new roles for a net gain of 78 million, the figure that gets cited in executive presentations. What gets cited less is that 39% of existing workplace skills will be disrupted or outdated within five years, and the churn that number implies is enormous even when the totals look balanced on a slide. Goldman Sachs Research narrows the lens to the US: 6 to 7% of the American workforce faces displacement under baseline AI adoption, concentrated in programmers, accountants, legal assistants, and customer service representatives, and it isn't only projection. Challenger, Gray & Christmas recorded nearly 55,000 US job cuts in 2025 that employers attributed directly to AI, up from roughly 4,600 when the firm first started tracking the category in 2023.

Most leaders hear 'people are afraid of AI' and reach for training programmes and upskilling budgets, and skills development matters. But it misses the deeper issue. The fear runs to identity, not skills. I recently spoke with a senior leader at a banking organisation of more than 40,000 people who told me their leaders, not the frontline, were the ones most afraid: afraid their expertise, judgement, and years of hard-won knowledge no longer mattered. I call this Identity Drift, the disorienting experience of watching a professional identity loosen as the capabilities a career was built on become automatable. Grounded in Role Identity Theory, it reflects that people don't just perform tasks, they build a sense of self around the expertise those tasks require, which is what makes this shift from past technology transitions different, and it's happening everywhere, often silently.

The mechanism behind that fear is the same one I've traced elsewhere. A 2025 study found that AI adoption erodes psychological safety, and eroded safety increases depression, but leadership changes the equation: when leaders were transparent and showed genuine concern during the transition, the negative effect on safety dropped significantly. The American Psychological Association's 2024 Work in America survey of more than 2,000 employed US adults fills in the picture: workers worried about AI taking over their job duties were significantly more likely to feel tense during the workday (54% against 36%), to say their work environment harms their mental health (47% against 28%), and to report they aren't learning things that will help their future (50% against 31%). They were also nearly twice as likely to say they plan to look for a new job within the year.

When psychological safety collapses, the fear goes underground rather than away. It shows up as Silent Resistance, where employees nod along in town halls, complete the mandatory training, and quietly go back to doing things the old way because engaging with the new tools feels like speeding up their own obsolescence. Underneath that often sits AI Shame, the shame of not understanding a technology everyone around you seems to have mastered, and the shame of admitting you used it because that signals your work is less earned. Forty-nine percent of employees are already concealing their AI use at work, and the hiding runs to the top, where 53% of leaders do the same.

The most powerful thing a leader can do is name the fear as rational rather than rushing past it with reassurance. Before launching another training initiative, help people answer a more fundamental question: who am I in this new world? That reframe isn't a communications task, it's a change management task, and it requires deliberate attention to Identity Drift before adoption can stick. Most organisations track licence activations. Psychological Safety is one of five drivers in the Organisational Adoption Profile, and it functions as the foundation the other four can't activate without. Yes, AI will displace roles, the WEF, Goldman Sachs, and Challenger's layoff data are all clear on that and it's accelerating. But the question that actually determines outcomes isn't whether AI will change work, it's whether an organisation is psychologically ready to move through that change without losing the people, trust, and institutional knowledge no algorithm can replicate.