Live-streaming in the context of the gaming community (i.e., live-streams of video game play) has become a popular form of entertainment, enabling real-time interactions between streamers and viewers. However, many of these interactions actively rely on language comprehension and production, therefore limiting the reach of any stream to a more restricted audience. Live-streaming being a very different type of media compared to other existing media, the effects and outcomes of translation methods traditionally used in video media are unknown. This study investigates how subtitles and AI generated dubbed voices as a form of machine translation affect viewer engagement and parasocial relationships (PSR) in viewing live-streams, by using mock videos imitating live-streams. Results suggest that translation helped the viewers to have a deeper understanding of the content regardless of language, compared to conditions without language comprehension support. Results have also suggest that language comprehension with machine translation leads to higher engagement and higher PSR, although the effects of each translation modality differed depending on the game titles. These findings highlight the potential for machine translation to be used in live-streaming, while also highlighting the potentially differential effects of modalities of translation across types of streams.
