Ninety-Four Percent Is Not a Rounding Error
The Bank of Korea spent four years watching what happened to young workers after ChatGPT launched, and this week it published the number: between June 2022 and June 2026, South Koreans aged 15 to 29 lost 285,000 jobs. Of those, 268,000 — 94 percent — vanished from sectors the central bank classifies as highly exposed to AI. Information services employment for young workers fell 31.4 percent. Publishing dropped 27.4 percent. Computer programming, systems integration, and management fell 16.6 percent.
Even now, somewhere, a strategy deck is calling this "efficiency."
The Jobs Didn't Disappear. They Changed Owners.
Here is the detail that turns a sad statistic into a genuinely uncomfortable one: while young workers were losing those 268,000 positions, workers in their 50s gained 230,000 jobs over the same four years — and 173,000 of those, 75.2 percent, were in the very same AI-exposed industries the young were being pushed out of. The jobs did not evaporate. They changed hands, and the hands that received them were, on average, three decades older than the ones that lost them.
Sources within the Central Banking Community don't typically editorialize, but the chart does it for them.
A Central Bank Official Explains the Obvious, Carefully
Oh Sam-il, who heads the Bank of Korea's Labor Market Research Team, offered the plainest possible summary: AI primarily performs tasks that are laid out in manuals, which means it can replace younger workers — the ones most likely to be doing exactly that kind of standardized, entry-level, manual-followable work. The report drew a formal distinction between "automation," where AI replaces a task outright, and "augmentation," where it merely assists a worker who keeps their job. One presumes the distinction is cold comfort to whoever landed on the wrong side of it.
The report also found that four-year college graduates have faced 7 percent average unemployment since ChatGPT's debut, versus 8.2 percent in the years before it — a number that looks like improvement until you remember it excludes everyone who left the labor market entirely rather than stay unemployed in it.
With the candor of an actuary reading a life-insurance table, the Bank of Korea has confirmed what younger workers already suspected: the manual said AI would help everyone. The data says it helped whoever was already established enough to be worth augmenting.
The Ladder Didn't Break. It Was Repositioned
Monthly outflow from these sectors among young workers rose roughly 32 percent, while inflow of new young hires fell about 11 percent — the classic signature of a door that swings only one way. In a development that will surprise no one who has been paying attention, the institutions measuring AI's economic impact keep arriving at the same finding by different routes: the technology does not eliminate work so much as redistribute it toward whoever already has a foothold, at the direct expense of whoever is still trying to get one.
The career ladder, as it turns out, was never removed. It was just quietly repositioned to start at fifty.
Sources: Korea JoongAng Daily · Xinhua
