Does generative AI prevent juniors from getting entry-level work?

The paper asks a collective-level question that individual productivity studies leave open: "if entry-level work is no longer reaching junior engineers, who will become the next generation of seniors?" Based on 14 semi-structured interviews in South Korea, with juniors at the threshold of entering software engineering and with senior engineers, analyzed by Reflexive Thematic Analysis, it names a foundational pattern of "Absorption": GenAI "redirects entry-level work into senior-AI workflows." The introduction sets the backdrop with figures it cites from other work, including a 25% fall in entry-level postings at the fifteen largest U.S. technology firms between 2023 and 2024 and Korean firms largely suspending open junior recruitment.

Three consequences follow from Absorption in the authors' account. Juniors lose "the productive struggle through which expertise once developed." That loss is then reproduced structurally through the collective normalization of GenAI use in university classrooms. And a perceptual asymmetry between seniors and juniors means neither side can correct the dynamic alone. The discussion sharpens the split: seniors describe GenAI taking over the entry-level tasks that once supported junior growth, while juniors describe GenAI use at university hindering the very skills that seniors, now working with GenAI, think juniors should have. The gap between the two groups widens from both directions.

The authors also argue that GenAI did not create the weakness it exploits. Seniors described their own development not as formal training but as "the incremental accumulation of hands-on work" that was never systematized, so GenAI has been "added onto" an already unprotected pathway. This moves the finding away from a simple deskilling story. The struggle was valuable, but nothing in the organization was built to guarantee it.

Against the nearest notes, this paper changes the level of analysis. Does AI assistance actually harm the way developers learn? and Does AI assistance help workers learn lasting skills? show an individual learner or worker who has the task in front of them and uses AI badly or well. Absorption describes the case where the task never reaches the junior at all. It is consistent with When does AI actually boost worker productivity?: the workflows that absorb the work belong to people who already hold the skills, and the paper's own concession is that individual productivity gains are real while the pipeline cost is a separate matter.

The excerpt does not say how many participants were juniors and how many were seniors. It reports no measures of junior competence, so the loss of productive struggle rests on what interviewees said, not on observed skill. The hiring statistics are cited from other authors, and the excerpt does not claim GenAI caused them. It also gives no recommendations and no evidence from outside South Korea. The implication that follows is modest. Studies of AI and learning should ask whether juniors are given the work at all, not only how they use AI on it, and the seniors-versus-juniors perception gap is a reason to check each side's view of the other's needs before designing training.