This is probably the topic with the widest gap between what gets published and what the data shows. Worth separating the two before making headcount decisions.
First: there is no collapse
The research published in 2026 agrees on something uncomfortable for both sides of the debate. The New York Fed’s job postings analysis finds no evidence of widespread displacement, and notes that AI is not the main driver of the hiring slowdown.
Anyone telling you AI is destroying employment at a macro scale is running ahead of the evidence.
Second: there is an effect, and it is very specific
Canaries in the Coal Mine?, from Stanford’s Digital Economy Lab and published in August 2026, finds something fairly sharp:
- Employment of workers aged 22 to 25 in AI-exposed occupations sits 19% below where it would be had it kept pace with their less-exposed peers.
- Those aged 22–25 in the most exposed occupations saw a 16% relative decline after ChatGPT’s release.
- Experienced workers show no comparable gap.
And the mechanism matters: the divergence operates through reduced hiring of young workers, not increased separations. Nobody is being let go. The entry door is being used less.
Third: the junior role is being redefined
PwC’s 2026 Global AI Jobs Barometer carries the finding that gives us most pause: AI-exposed entry-level roles are seven times more likely to demand traditionally senior skills — judgement, leadership, discernment. And those roles have grown 35% since 2019, while other entry-level roles fell 10%.
Entry-level work is not disappearing. The bar to enter is rising.
What this means if you run a company
Beware the short-term saving that leaves you with no bench. If you stop hiring juniors because AI covers their tasks, in five years you will have no seniors. Junior work was never only output: it was the mechanism by which someone learned the craft.
A junior’s value moves. It stops being “does the repetitive work” and becomes “reviews, challenges, and spots when the system’s output is wrong”. That requires training them differently and earlier.
Just over a quarter of employers say AI has reduced the need for tasks juniors performed — but more than half are already having that conversation internally. The gap between those two figures is what happens over the next two years.
What we recommend
- Don’t cut entry-level hiring on a projection. Cut it when you have the measurement, not when you have the slide.
- Redesign the junior role around verification. Someone who can spot a system error is worth more than someone who produces the draft.
- Train judgement, not tools. The tool changes every six months; knowing when to distrust it does not.
- Document tacit knowledge now. If your senior staff retire and you trained no successors, AI will not save you: it cannot learn what was never written down.
The honest conclusion is that the effect is real, measured, and far narrower and more specific than the public debate suggests. Being narrow does not make it any less serious for the people in that band.
Sources
- Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of AI — Stanford Digital Economy Lab (PDF)
- Do Job Postings Show Early Labor-Market Effects of AI? — Liberty Street Economics, Federal Reserve Bank of New York
- PwC 2026 Global AI Jobs Barometer — PwC
- Labor market impacts of AI: A new measure — Anthropic
- The Impact of AI on the Early-career Labor Market — NACE