The research team approached the problem not by trying to pin down the causes of declining recruitment, but by examining how the roles of junior developers and the process of building professional expertise are actually changing on the ground. They chose South Korea as a case study in which these shifts appear in an early and highly visible form. According to data cited in the paper, AI is involved in 51.8 percent of work-related activities in Korea, nearly twice the comparable figure for the United States, while conventional entry-level hiring has contracted sharply. Entry-level job postings at the fifteen largest U.S. technology companies fell by 25 percent between 2023 and 2024, and IT job postings in Korea declined by 43 percent over the same period, with positions targeting entry-level candidates accounting for just 4.4 percent of all postings. The researchers caution that these declines cannot be attributed solely to generative AI, since post-pandemic workforce adjustments and the economic slowdown have also played a role, but the structural transformation of junior work was unmistakable in their data.

At the heart of the study are semi-structured interviews, each lasting approximately fifty minutes, with fourteen participants: six senior developers currently working in the industry and eight junior participants preparing to enter software development careers. The senior participants were required to have at least six years of professional experience and to have entered the workforce before the emergence of generative AI, while the junior participants were undergraduates whose university years spanned both the pre-AI and post-AI eras. The team analyzed the transcripts using reflexive thematic analysis, a qualitative method in which researchers iteratively examine their own interpretive role while identifying recurring patterns across interviews. Four major themes emerged, and together they sketch a system in which the loss of junior-level work is inseparable from the loss of junior-level learning.

The first theme was the shift of junior-level tasks toward what the researchers describe as “senior plus AI” workflows. Senior developers increasingly use AI to directly handle basic implementation, debugging, and documentation that would previously have been assigned to junior staff, and in doing so they absorb the very tasks through which novices once gained hands-on experience. One startup founder interviewed for the study said the company had substantially reduced its junior workforce without any measurable impact on productivity, asking pointedly, “What makes a junior developer better than a KRW 100,000-per-month AI subscription?” The economic logic of that question is difficult to dispute in the short term, and it explains why individual firms have strong incentives to make the substitution. The study’s point is that the costs of this rational choice do not disappear; they are displaced onto a different group of people at a different point in time.

That displacement is captured in the second theme: what is disappearing is not simply work, but the “opportunity to fail.” The researchers frame this through the concept of “productive struggle,” the well-documented principle in learning science that when learners make mistakes and correct them on their own, they develop not only knowledge but also the metacognitive ability to recognize what they do not know and to detect errors in the results they produce. Junior participants reported that they could now achieve good results with far less effort than before, yet many described a troubling sense of “not knowing what I don’t know.” One participant who took two courses in the same field received the same grade in both, but said that one course left them with actual knowledge while the other left them only with “the ability to use AI.” The single difference between the courses was whether the assignments could be completed using AI. The researchers found that current educational assessment methods do not adequately capture this distinction, meaning grades can mask profound differences in what students have actually learned.

Senior developers voiced the same concern from the other side of the career ladder. One participant explained, “The reason you become a senior is that experience teaches you what not to do. Now juniors can no longer learn from bad examples.” This observation connects directly to the study’s theoretical grounding in “situated cognition,” the idea that the problems a person can perceive depend on the position they occupy. Senior developers and the organizations with the power to address the problem may find it difficult to fully appreciate the circumstances facing juniors, while juniors who directly experience the problem lack the power to change educational and hiring structures. The asymmetry is captured starkly in the words of another senior participant: “We have 20 years of accumulated experience, so we can judge whether AI outputs are right or wrong. The next generation will not be able to reach that position. So we are fine.”

The third and fourth themes underscore that these dynamics are structural rather than matters of individual choice, and that they will not correct themselves. When every one of a student’s peers uses generative AI, grading on a curve effectively becomes a mechanism that compels AI use, since opting out means accepting a competitive disadvantage. In the courses taken by the interview participants, AI use was rarely restricted, so the loss of failure-based learning was built into university classrooms just as surely as into workplaces. Meanwhile, because seniors and juniors perceive the same situation so differently, no natural feedback loop exists to trigger correction. The researchers also found that the problem did not originate with generative AI alone: senior developers themselves largely built their expertise through hands-on work and trial and error rather than through formal training systems. In this sense, generative AI has not eliminated a newly created development pathway; it has accelerated the erosion of a pathway that was never institutionally protected in the first place.

The study argues that the solution lies in deliberate institutional protection rather than individual effort, and it draws on examples from other high-risk fields where automation has already forced this reckoning. In aviation, the Federal Aviation Administration’s SAFO 13002 advises pilots to maintain opportunities for manual flying so that automation does not erode their manual flying skills, and the nuclear industry requires regular simulator retraining to keep operators’ response capabilities sharp. These are examples of industries deliberately preserving the experiences necessary for developing and maintaining expertise even when machines can perform the underlying tasks. The research team proposes analogous reforms across three areas of software engineering. At universities, courses in which AI cannot achieve the learning objectives on behalf of students should be designated as required, and “reducing AI dependence” should serve as a criterion for evaluating educational quality. In hiring, employers should evaluate not only candidates’ ability to use AI to produce results quickly, but also their ability to recognize gaps in their own knowledge and detect errors while working alongside AI. Within companies, learning opportunities should be deliberately created for entry-level developers, for example by assigning them small modifications to actual products and letting them experience the full process from commits and code review through deployment.

Professor Moon, who supervised the research, framed the findings as a warning about a hidden ledger. “Some of the productivity we are gaining today may effectively be borrowed in advance from the expertise of the next generation,” he said. “Because the cost is borne by a different group and at a different point in time from those making the decisions, it is structurally difficult to recognize. This study is significant because it empirically reveals that blind spot.” He added that, as a department that develops AI technologies itself, SNU bears a particular responsibility to ask what those technologies are doing to the development of the next generation, and that the team plans follow-up research focused on university education and on incorporating the findings into the design of the undergraduate curriculum. The paper’s first author, Ph.D. student Sumin Yu of SNU’s M.IN.D Lab, conducts empirical research on how generative AI affects human learning, decision-making, and social institutions, and plans to continue studying the interaction between AI technologies and society across a range of fields.

The significance of the work extends beyond software engineering. AIES, jointly organized by the Association for the Advancement of Artificial Intelligence and the Association for Computing Machinery, is regarded alongside ACM FAccT as one of the two leading venues for research on the ethical and societal impacts of AI, and the ninth edition of the conference will be held in Malmö, Sweden, from October 12 to 14. The research was supported by the National Research Foundation of Korea, the Institute of Information & Communications Technology Planning & Evaluation, and the BK21 FOUR Education and Research Program at Seoul National University. What the study ultimately documents is a paradox at the center of the AI productivity boom: the judgment that makes senior engineers valuable is produced by exactly the kind of slow, error-prone, hands-on work that AI now performs so cheaply. An industry that optimizes away that work may find, a decade from now, that it has optimized away its own future experts, and the bill for today’s efficiency gains will come due in the form of a generation that never learned what it does not know.

Subject of Research: How generative AI erodes the professional development pathway of junior software developers

Article Title: SNU professor Taesup Moon’s team reveals how generative AI is eroding the development pathway for software developers

Article References: SNU professor Taesup Moon’s team reveals how generative AI is eroding the development pathway for software developers. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: generative AI, software engineering, junior developers, skill development, productive struggle, situated cognition, AIES 2026, Seoul National University, entry-level hiring, AI ethics, developer education, automation

The study rests on a theoretical foundation drawn from three pillars of learning science. The first is cognitive load theory, developed by John Sweller, which describes how sensory memory, working memory, and long-term memory interact when learners process new information. According to this framework, the difficulty of learning a programming paradigm depends on the level of abstraction and the number of interacting elements a learner must juggle simultaneously. When a developer trained in object-oriented thinking encounters functional programming, the unfamiliar constructs must be integrated with existing mental schemas, a process that can easily overload working memory and slow early progress. The second pillar is transfer learning theory, which explains how skills acquired in one domain can either help or hinder learning in another. Prior experience with object-oriented interfaces may ease the adoption of similar constructs in Java, for example, while deeply ingrained imperative habits can actively interfere with understanding functional programming’s insistence on immutability. The third framework, expertise development theory, describes how knowledge accretes over years of practice, with rapid early gains followed by plateaus as learners refine their conceptual understanding, a pattern that maps directly onto the classic shape of a learning curve.

What makes programming paradigms such a compelling subject for cognitive research is that they are not merely syntactic conventions. Each paradigm encodes a distinct cognitive organization of the problem space. Object-oriented programming encourages learners to model systems as networks of classes and objects that communicate through message-passing, emphasizing encapsulation and modular design. Functional programming demands immutability, higher-order functions, and declarative reasoning that resembles mathematics more than step-by-step instruction. Logic and other declarative paradigms ask learners to specify desired outcomes rather than explicit control flows. Because these mental models differ so profoundly, switching paradigms forces developers to reorganize how they represent problems internally. The study highlights a phenomenon the literature calls paradigm adherence or transfer interference: developers anchored to their first-learned paradigm tend to reach for the wrong tools when working in another, searching for solutions based on residual knowledge from a previous approach. This not only impedes comprehension but can measurably slow the rate of learning, which is precisely why the author argues that paradigm acquisition is neither uniform nor linear and deserves empirical measurement in its own right.

To capture these trajectories, the researcher adopted a cross-sectional, survey-based design combining quantitative and qualitative methods. A structured online questionnaire collected data on years of experience per paradigm, self-rated proficiency at six time points from the start of learning through the present, perceived difficulty on a five-point scale, motivation, and hours of study. Seventeen developers responded, a small but diverse group that included students, researchers and academics, and professional developers, with bachelor’s degree holders making up roughly 52.9 percent of respondents. The questionnaire was content-validated by two academic researchers in computing education and piloted with two developers before distribution. Responses were anonymized, exported, and analyzed in Python, with descriptive statistics, Pearson correlations, and a paired t-test forming the quantitative core, while open-ended responses were subjected to a six-phase thematic analysis following the framework of Braun and Clarke. The author is candid about the limitations: with such a modest sample and retrospective self-reports, the results are descriptive and hypothesis-generating rather than generalizable, but they offer a structured operationalization of paradigm learning that larger studies can build upon.

The experience data revealed a pronounced hierarchy among the paradigms. Object-oriented programming showed the highest mean prior experience at approximately 3.15 years, followed by functional programming at about 2.32 years and procedural programming at roughly 2.21 years. At the bottom sat the reactive/event-driven and concurrent/parallel paradigms, with mean experience of only about 1.82 and 1.47 years respectively. This distribution suggests that object-oriented and procedural styles remain the default foundation of most developers’ education, while reactive and concurrent approaches are treated as specialized or advanced domains that many practitioners simply never encounter in depth. The pattern, the study argues, reflects curricular biases and structural barriers in computing education rather than any inherent order in which paradigms should be learned, and it hints at a potential misalignment between what universities teach and what the modern software industry increasingly demands.

The learning curves themselves told a story of steady, if effortful, progress. Average self-rated proficiency began at a baseline of roughly 3.71, climbed to about 4.0 by the three-month mark, and reached approximately 5.18 by the time of the survey. A paired t-test comparing starting and current proficiency produced a statistically significant result, with t equal to negative 4.47 and p below 0.05, indicating a measurable learning gain across the group. Difficulty ratings, however, skewed heavily toward the difficult end of the scale, with only a minority of participants describing the learning process as easy or moderate. The author attributes this to two non-exclusive factors: the intrinsic complexity and abstraction of the paradigms themselves, and the particular demographic composition of the sample, which included many learners still embedded in academic settings. Correlation analysis added further texture, showing a positive relationship of about 0.45 between general experience and current proficiency, positive correlations between experience in related paradigms such as functional and object-oriented programming that suggest foundational knowledge transfer, and a positive link between total hours invested and final proficiency.

Perhaps the most vivid findings came from the qualitative strand. Thematic analysis of the open-ended responses surfaced three dominant motivators for learning a new paradigm: academic and curricular drivers such as university courses and degree requirements, employment and project needs including job requirements and freelance work, and intrinsic motivation rooted in personal interest and self-growth. Participants who learned with a definitive practical goal in mind, such as capacity building or project demands, reported the highest average proficiency levels, reaching up to 7.0 on the self-rating scale. One participant observed that most paradigms are intertwined, noting that understanding the basics of procedural and object-oriented programming makes learning other paradigms easier. Another described a slow learning trajectory attributed to self-directed study, while a third reported learning far more after joining a team, underscoring the value of collaborative and structured support. A word cloud of recurring terms placed academic and course above all others, with job, project, and personal appearing as secondary drivers.

Taken together, the results paint a picture of developers leaning heavily on self-directed learning, including documentation, tutorials, and hands-on experimentation, to acquire the more advanced paradigms that formal instruction leaves uncovered. The study frames this as a quiet failure of the traditional educational model: learners are well prepared for conventional tasks built on procedural and object-oriented foundations, but the cognitive strain of abstract or concurrency-focused paradigms, combined with limited early exposure and weak pedagogical scaffolding, slows their progression precisely where industry needs them to be fluent. The author recommends diversifying paradigm exposure at early learning stages, providing more robust scaffolding for abstract concepts, and integrating real-world, project-driven learning into curricula. The findings also connect to broader educational research suggesting that structured bridging activities, critical thinking exercises, and algorithmic problem-solving approaches can shorten learning curves in programming and related domains.

The study’s limitations are acknowledged plainly and are worth keeping in view. Seventeen respondents cannot represent the global developer population, retrospective self-assessments capture perceived rather than actual learning trajectories, and the cross-sectional design rules out causal inference about how proficiency develops over time. Recall bias and subjective interpretations of proficiency are inherent risks of the method. The author calls for longitudinal designs that track real proficiency changes, larger and more diverse samples spanning multiple institutions and regions, and experimental studies evaluating specific interventions such as visualization tools, AI-assisted learning systems, and project-based modules. Even so, the work makes a genuine contribution by treating paradigm acquisition as a measurable cognitive process with comparable learning curves, rather than a vague matter of pedagogical taste. As software systems grow more concurrent, reactive, and distributed, the gap between what computing curricula emphasize and what developers must master is only likely to widen, and studies like this one provide the empirical footing needed to close it.

Subject of Research: Empirical study of learning curves and how developers acquire proficiency across five programming paradigms

Article Title: Learning curves and programming paradigm acquisition in developer education

Article References: Learning curves and programming paradigm acquisition in developer education. (n.d.). https://doi.org/10.1007/s44217-026-02124-2

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02124-2

Keywords: programming paradigms, learning curves, developer education, cognitive load theory, object-oriented programming, functional programming, concurrent programming, computing curricula, self-directed learning, transfer of learning, thematic analysis, software engineering education