Generative AI for Video Translation: Consumer Evaluation in International Markets

Abstract

Generative AI tools (e.g., HeyGen, Adobe Firefly, Invideo AI) now enable marketers to translate videos not only by converting language but also by adjusting speech style, voice, and lip movements. Following this advancement, this exploratory study examined differences in perceived translation quality between AI-translated and human-translated marketing videos in international contexts. Two between-subjects experiments were conducted, involving English-to-Indonesian translation (Study 1) and Indonesian-to-English translation (Study 2). AI translation consistently yielded lower perceived naturality and accent neutrality than human translation. For language comprehension, AI performed worse in Study 1 but better in Study 2, indicating that translation direction matters. However, despite the perceptual differences, the two translation methods did not affect customer engagement intention. This study offers early evidence on how consumers evaluate AI video translation and provides 12 directions for future research.