본문 바로가기
  • Home

Generative AI in audiovisual translation education: A case study of Chinese-Korean subtitling

  • The Journal of Translation Studies
  • Abbr : JTS
  • 2026, 27(1), pp.355~381
  • DOI : 10.15749/jts.2026.27.1.011
  • Publisher : The Korean Association for Translation Studies
  • Research Area : Humanities > Interpretation and Translation Studies
  • Received : January 29, 2026
  • Accepted : March 16, 2026
  • Published : March 31, 2026

KIM, HAE RHIM 1

1울산대학교

Accredited

ABSTRACT

This study designs a Chinese-Korean audiovisual translation course for undergraduate students using DeepSeek. Using the full transcripts of student-DeepSeek interactions submitted as translation assignments, the study examines how students construct prompts during the translation process and how those prompts are reflected in DeepSeek’s outputs. The analysis reveals that students generally relied on one-directional, command-based prompting and failed to fully utilize DeepSeek’s interactive conversational capabilities. Given that subtitling involves strategic decision-making and creative rewriting under the multimodal and spatio-temporal constraints of audiovisual media—areas in which DeepSeek demonstrates clear limitations—the findings underscore the need for systematic instruction that foregrounds human translator agency. Specifically, translators must be positioned as active agents who refine translation outcomes through iterative feedback and critical engagement with AI-generated outputs. The analysis also shows that students tend to uncritically accept inappropriate or problematic outputs without recognizing their limitations. Accordingly, future generative AI-based translation education should incorporate source-target text comparison and verification into its design., with a particular focus on fostering students’ critical awareness of hallucinations, omissions, and ethical issues in AI-generated translations.

Citation status

* References for papers published after 2024 are currently being built.