The scariest prediction about artificial intelligence was always the simplest: machines get good enough at knowledge work, companies cut payrolls, and millions of educated professionals find themselves competing for a shrinking pool of jobs. On Sunday, Workday released its October Global Workforce Report with a headline built to reassure: AI is rewriting jobs more than it is cutting them. Just 28% of business leaders expect AI to reduce headcount, while 40% expect it to squeeze more out of the staff they already have. Pair that with a strong August US jobs report, 162,000 positions added against 4.1% unemployment, and the apocalypse looks postponed. The trouble is that "postponed" and "not happening" are not the same thing, and the same reports quietly explain why.
Start with what the reassuring number hides. Workday found that the ordinary machinery of getting ahead is seizing up. Internal moves to new roles fell at 57% of employers year over year, promotion rates stayed flat, and nearly four in ten workers said their company reorganised in the past year. Your job may be safe, in other words, while the path out of it quietly closes. The skills employers reward are churning fast too: demand for basic AI skills like prompting peaked in January 2026 and then dropped 25%, while demand for people who can actually build AI tools rose 51%. Knowing how to ask a chatbot a question is already a depreciating asset.
The aggregate figures mask the real action, which is happening inside industries rather than across them. The Economist's recent analysis, summarised by Linkdood, argues the feared collapse has given way to an AI-linked boom, but a lopsided one. Industries tied to the data-center build-out have added roughly 320,000 more jobs than construction and manufacturing trends would predict since 2023; professional services closest to AI have added around 730,000 excess jobs. Meanwhile the US Bureau of Labor Statistics projects office and administrative support roles will shrink by about 752,000 between 2025 and 2035, the largest decline of any occupational group. A firm can cut 500 junior analysts and hire 500 AI engineers and data-center technicians, and the national total barely moves. The displaced workers do not experience a total; they experience a door closing.
That reshuffle lands hardest on the people with the least footing. Reporting from India, where IT services employ millions, captures the squeeze from the other side: firms there face weak deal flow and clients demanding that AI's cost savings be passed back as lower prices, even as basic coding, data entry, and customer support become automatable. The worry voiced across those accounts is the same one echoing in Western economies: if AI does the tasks once handed to juniors, young workers lose the rungs they used to climb. One Indian commentary reached for a nineteenth-century analogy, the "Engels' Pause," when Britain's industrial output soared for decades before ordinary wages caught up. Productivity can boom while workers wait, sometimes a very long time, for their share.
So the honest reading of this week's data is neither panic nor relief. AI is not emptying the offices; it is rewiring who sits in them and how they got there. That is arguably harder to govern than mass layoffs, because it is diffuse and slow and shows up as a stalled promotion or a vanished entry-level posting rather than a headline redundancy. A Newsweek essay added a quieter warning worth keeping in view: the deeper risk may not be losing the job at all, but leaning on the machine until the judgment that made you good at it quietly atrophies. The jobs are mostly still there. Whether the ladder, and the skill, survive the rewrite is the open question.