This study examines how employment signals in the Korean labor market were rearranged across worker groups after November 2022, when ChatGPT emerged as a major technological shock. Focusing on new labor market entrants and incumbent workers, we use news texts as proxy indicators to complement the lagging nature of administrative employment data. Based on 112,600 AI and employment-related news articles, we apply a deep learning-based topic modeling pipeline and a frame-level two-way fixed effects difference-in-differences design. The results show that media signals after AI diffusion did not converge into a simple narrative of job destruction. Instead, heterogeneous signal shifts appeared across groups and frames. Macroeconomic and corporate performance frames were relatively strengthened, whereas signals related to new entrants weakened within the labor restructuring frame. These findings suggest that firms may respond to technological uncertainty by limiting external hiring and prioritizing internal workforce redeployment.
목차
Abstract 서론 문헌 연구 연구 방법 데이터 및 전처리 처리집단 정의 실증 분석 모형 연구 결과 결론 연구결과 요약 및 결론 시사점 연구의 한계 및 향후 연구 방향 참고 문헌