
21 Sep 2026
AI-adopting firms add jobs — mostly for seniors, not juniors
Companies that adopt AI are adding more jobs overall than peers that don’t — but a new Stanford–King’s College London paper released Monday finds those gains skew hard toward senior workers while junior employment falls.
JOBS desk — the headline that AI creates jobs hides a mix change: the first rungs shrink while the experienced seats grow.
The headcount change, as the authors told that wire. At companies adopting AI, the number of employees in senior positions rose 6.7% over five years. Junior employment fell 3% in the same period. Senior and junior are the paper’s labels for the higher rung and the lower rung. The wire does not print the exact job-title cut, so. A percent here is a change in the number of people, not a change in their share of the staff.
Overall hiring at those firms still rose. Even so, the share of the staff who are junior fell by 1.9 percentage points. A percentage point is one step on the share itself. If juniors were 20% of the staff and then 18.1%, that is a drop of 1.9 points, even if the building hired more people. A firm can add jobs and still end up with a smaller slice of those jobs held by juniors, when the new seats go to seniors.
The shift away from lower-level workers showed up in numerous countries. The wire names Brazil, Saudi Arabia, and the UK. It is not a claim that the pattern is the same size in every one of the 41 countries, and it is not a ranking this desk built.
The authors’ line, quoted on the wire: “AI is labor saving for junior workers and labor expanding for seniors in exposed occupations.” Labor saving means the work needs fewer of those people. Labor expanding means it needs more. An exposed occupation is a job where AI can take on a large share of the tasks. They also wrote that junior employment losses run somewhat deeper in richer and more digitized economies. Digitized, here, means more of the work already runs on computers and software. They are not a wage study.
The sample, still the wire. The academics analyzed 1.25 billion job postings and 154 million employment records, from January 2021 through March 2026. A job posting is an ad for an open role. An employment record is a count of someone actually in a job. Postings show what firms were trying to hire. Records show who was on the payroll.
Computer and math jobs moved too, and not as a junior boom. Roles the researchers call computer-related and mathematical, which they place among the most exposed to AI, rose by 0.8 percentage points at companies that adopted AI. That 0.8 is a share, the same kind of number as the 1.9-point junior drop, not a 0.8% change in headcount. There, too, the shift was toward senior workers. The wire does not say how large that senior tilt was inside computer and math. A rise in computer and math jobs is not a rise in junior computer and math jobs.
A separate note, and not a result of Monday’s paper. The same dispatch says Chandar and colleagues at the Stanford Digital Economy Lab wrote last month that employment among 22- to 25-year-olds in jobs highly exposed to AI is 19% below the path it would have taken if it had kept pace with less-exposed roles. The wire’s sentence does not name a country for that 19%.
The wire’s own lead is that AI-adopting firms are adding more jobs than firms that are not. The wire prints none. The Insurance Journal page credits an October 2024 job-fair picture. That picture is not the paper. The card on this filing is a claim diagram of the senior +6.7% and junior −3% figures.
Plain English for the rest of the card: senior / junior = the paper’s higher rung and lower rung; the wire does not print the title cut. percentage point = one step on a share, so a 1.9-point drop is not the same thing as a 1.9% drop in headcount. exposed occupation = a job where AI can take on a large share of the tasks. digitized economy = a country where more of the work already runs on computers. job posting = an ad for an open role. employment record = a count of someone in a job. This filing is the Monday paper as Bloomberg and Insurance Journal reported it. It is not the PDF and not a wage table.
RELATED
On 21 Sep 2026, Bloomberg reported that companies using artificial intelligence are adding more jobs than companies that are not. The same dispatch, by Saijel Kishan, ran the same day on Insurance Journal. The gains favor senior employees over junior workers, in a new study covering 41 countries. The study is a paper by Bharat Chandar of Stanford University and Bouke Klein Teeselink of King’s College London, released Monday. Artificial intelligence, or AI, here means software that can take on tasks people were doing at work. These lines are the wire’s account of the paper.




















