생성형 AI 기반 번역 교육의 창조성 함양 효과: 광고 카피의 시각적 트랜스크리에이션을 중심으로

Abstract

This study explores a transcreation-centered model of translator education, in which undergraduate students taking a practical translation course transform advertising copy into visual content using a multimodal generative AI tool. Contrary to the view that generative AI threatens human creativity, the findings of this study highlight the potential of AI as a creative collaborator that stimulates imaginative thinking and supports human agency in translation. By analyzing both the process and outcomes of the transcreation tasks, the study identifies key patterns in how students constructed prompts and intervened in image generation. Specifically, the process analysis reveals the following patterns. First, students designed prompts based on their interpretation of the original advertisement and its unique selling proposition, using rational appeals, emotional appeals, or both. Second, they actively shaped their own advertising concepts using imaginative storytelling or stylistic instructions to reflect personal creativity. Third, they fine-tuned visual elements such as color, lighting, and composition to enhance the aesthetic coherence between text and image. In terms of outcomes, the study identifies three main types of human intervention by comparing initial AI-generated images with students’ final selections. The first involved proactive concept-setting and collaboration with AI, where students took a leading role. The second showed a gradual development of creative direction through iterative engagement with AI outputs. The third entailed selective refinement of AI-generated content, reflecting a responsive rather than passive mode of collaboration. Overall, this study demonstrates how AI-supported transcreation tasks can foster creative autonomy in translation education and lays the groundwork for pedagogical models in which human translators and AI engage in dynamic, mutually enriching collaboration.