Discourse in Human-AI Collaboration: A Pragmatic and CDA Study on Translating Linguistically Unique Indonesian Songs Across Generational Audiences

Abstract

This study examines how discourse, context, and generational orientation shape AI-assisted translation, focusing on two Indonesian songs with unique linguistic features: “Stecu Stecu” and “Pica Pica.” Using a qualitative-descriptive case study approach, the research combines pragmatic linguistic theory, as outlined by Geoffrey Leech, with Fairclough’s three-dimensional model of Critical Discourse Analysis (CDA). ChatGPT was tasked with translating the songs into English using prompts tailored for Millennial, Generation Z, and Generation Alpha audiences. Findings show that AI can adjust tone, vocabulary, and style according to generational characteristics: Millennial outputs are smooth and clear, Gen Z versions use slang and digital expressions, while Gen Alpha translations are simple, playful, and easy to understand. These results suggest that AI can reflect audience-specific language use but also demonstrate that AI-generated translations are shaped by underlying social values and are not entirely neutral. The study underscores the importance of context-aware prompting and human oversight in AI translation, contributing to broader discussions in Translation Studies and responsible language technology use.