어휘적 결속성(Lexical Cohesion)이론을 적용한 생성형AI 번역 활동 연구 — 인간번역⋅신경망기계번역⋅챗GPT번역 비교를 중심으로
A Study of Generative AI Translation Activities Applying Lexical Cohesion Theory: A Comparative Analysis of Human Translation, Neural Machine Translation and ChatGPT Translation
- 발행
- 2026 등록KCI API에는 발행일 항목이 없어 원문 등록일을 쓰고 있습니다. 실측으로 발행보다 최대 6년 늦습니다 — 재수집하면 발행연월로 바뀝니다.
- 소속·발행
- 고려대학교
- 출처
- 국내 KCI
- DOI
- 10.26585/chlab.2026..91.002
- 원문
- 원문 보기 ↗
개념
키워드
Generative AI Translation, Text, Lexical Cohesion, Prompt Engineering, Generative AI Translation, Text, Lexical Cohesion, Prompt Engineering
초록
Recent advances in artificial intelligence (AI) have brought about profound transformations across many domains of human life. In academia—particularly in foreign language education—one area that demands systematic adaptation and investigation is AI-based translation, which has demonstrated remarkable growth since the emergence of Neural Machine Translation (NMT) in 2016. However, the operational paradigm of NMT primarily relies on one-to-one processing of parallel corpora at the lexical, sentential, and paragraph levels in order to secure equivalence. This processing framework reveals inherent limitations in capturing deep intra-textual mechanisms across global context and in integrating extra-linguistic factors into the generation of translation outputs.
 With the advent of generative AI–driven translation systems, exemplified by ChatGPT, a paradigm shift has occurred, raising the need for translators and educators to assume new roles in translation practice, pedagogy, and professional application. In response, this study proposes the necessity of developing specialized prompts for translation tasks and seeks its theoretical foundation in classical text-linguistic frameworks