How Does Generative AI Function as an Active Cross-cultural Communication Agent? A case study of Japanese literature translation using ChatGPT-4

Abstract

This study explores how generative AI, specifically ChatGPT-4, transcends its role as a mere translation tool to become an active agent in cross-cultural dialogue and meaning generation. Against the backdrop of the rapid evolution of AI translation technologies, this research employs a posthumanist framework, drawing on Karen Barad’s (2007) agential realism and Roland Barthes’ (1967) concept of the “death of the author,” to examine how AI can function as a dynamic mediator in translation processes. Using the Japanese novel Jimmy as a case study, the research highlights a dual-method approach: translating the original Japanese text into English and Chinese using ChatGPT-4 and conducting a transnational reading session in Penang, Malaysia. The novel, chosen for its accessibility and themes of cultural intersection, served as a medium to examine the potential of AI in fostering cross-cultural understanding. Participants from Japan and Malaysia engaged in discussions on key chapters, offering interpretations shaped by their unique cultural contexts. The findings reveal that generative AI, through mechanisms like critical reading question generation and knowledge base updates, enables translations to incorporate deeper cultural nuances and reflect dynamic interpretations. The process illustrates AI’s capacity to facilitate a “diffractive” translation approach, where intersecting perspectives generate new layers of meaning. However, limitations emerged, such as ChatGPT-4’s RAG filtering, which restricted certain translations and necessitated human intervention.
This interplay between AI and human translators underscores the evolving nature of their collaborative relationship. The study highlights generative AI’s potential to transform cultural dialogue and translation practices. By acting as a bridge between diverse cultural perspectives, AI can contribute to educational initiatives, enhance the quality of translations in the industry, and foster mutual understanding in cross-cultural interactions. Future research should explore the scalability of this approach across different languages and cultural contexts, further refining the collaborative dynamics between AI and human participants.