The back-and-forth among corporate executives and public figures over how much of people's work AI is taking has heated up again. But over the same period, reports came out that call center jobs had fallen sharply in Phoenix in the United States, and videos of young men knocking over or dragging away unmanned delivery robots in downtown Los Angeles circulated online.
OpenAI CEO Sam Altman said at a seminar in Sydney, Australia, on May 27 that "the end of jobs will not come." His point was that predictions of entry-level office jobs disappearing en masse had missed the mark. A labor market report released by Anthropic points in a similar direction. It says there was no clear change in the unemployment curve for existing white-collar workers even after AI was adopted, and that companies responded by cutting new hiring instead of laying people off.
The "Observed Exposure" measure Anthropic used gauges how much AI is actually used in a given job by observing it. It carries the concern that theoretical replaceability and real-world usage are different things. The scope this indicator measures is limited to the United States, and to the present moment.
Measuring Korea gives a different picture. In its February 2025 report "AI and the Korean Economy," the Bank of Korea analyzed that 51% of domestic jobs fall within the range affected by AI adoption, and that 27% face a high likelihood of being replaced or seeing incomes fall. The measured figures from the United States and Korea's estimate of potential impact are measuring different things, so they do not collide head-on.
It is a shock in the form of people already at work not being pushed out, while the door narrows for those trying to enter for the first time. In Korea as well, the number of young workers has continued on a downward trend while the total number of employed people has grown.
Built to travel at low speed and avoid people automatically, these machines cost several thousand dollars each. There is no basis for tying the motives of those who knocked them over into a single one, but it is clear that this happened at the point where machines are taking over deliveries that people used to make.
The discussion of how to respond is not moving as fast. Kim Ji-hyung, chairman of the Economic, Social and Labor Council, said in an interview at his office on Saemunan-ro in Jongno-gu, Seoul, on May 26 that responding to AI requires assuming even "the worst case" and preparing through social dialogue. Temporary mass unemployment and polarization should also be put among the things to prepare for. He saw the issue as not confined to particular occupations and spreading in every direction.
The bargaining frameworks that handle distribution are also being put to the test. With the recently concluded Samsung Electronics performance bonus negotiations as a trigger, there is talk that how to divide excess profits could rise onto the bargaining agenda at a number of companies in Korea. Who receives the increased share when productivity shifts toward machines is a part that employment statistics do not answer.
On June 24, the "5th People and Digital Forum-The Future of Work and Learning in the AI Era" will be held. Whether the talk exchanged there becomes an extension of the debate or a discussion that draws up countermeasures on the basis of measured data remains open.
