This is an opinion piece by Georgios Petropoulos in MIT Technology Review (26 May 2026) that summarises other studies. We therefore cite the studies as reported there and suggest checking the originals.
What the piece cites
- A Stanford Digital Economy Lab working paper (November 2025) reporting a 16% relative employment decline for workers aged 22 to 25 in the most AI-exposed occupations, with no similar decline for experienced workers in the same occupations.
- New York Fed data for Q4 2025 showing recent-graduate unemployment of 5.6% and underemployment of 42.5%.
The caution
The author notes that no single statistic proves AI is the sole cause, and that a broader hiring slowdown also affects young workers. The honest summary is that early signals are concerning and the causes are still being established.
Planning implication
If entry-level tasks are automated, organizations still need people who will become experienced. Plan how juniors learn: supervised work on real problems, rotations and explicit coaching, rather than assuming experience will appear.
Treat this as an open question, and avoid using it to justify cuts or to dismiss concerns.