Utilizing AI in game localization : challenges and benefits

Abstract

This thesis examines the challenges and benefits of AI-assisted translation tools on game localization processes. The research was conducted with a qualitative method. Three different sources were used to collect data: expert interviews, comparative translation samples, and observations and field notes. The data obtained from these sources were evaluated with thematic analysis method. This analysis was made with MAXQDA software, and the theme outputs were taken from this software. According to the findings, the most important advantages of AI-assisted translation tools include (1) saving time in technical, short and repetitive content, (2) ensuring terminological consistency in short texts, (3) increasing productivity and (4) effective hybrid models integrated into translation processes. However, some limitations of these tools are also noteworthy. Participants stated that AI is still inadequate, especially in creative content such as narrative, context-based and cultural texts. The main disadvantages include (1) lack of context and cultural adaptation, (2) inability to reflect emotional tone in creative texts, (3) inconsistency throughout the dense texts, and (4) decrease in human-oriented work. This study draws attention to the importance of hybrid models where artificial intelligence and human translators work together. It also shows that human-centered decisions will continue to be as decisive as technological developments in localization processes in the future.