Key points are not available for this paper at this time.
In the context of the rapidly evolving AI-generated translation landscape, this study explores the viewer experience of GPT-4-generated humor translations compared to human translations for dubbing purposes. Despite advances in AI technology, challenges persist in capturing the nuances of humor and cultural references, which are essential for effective humor translation. A group of cinephiles was provided with humor dialogues translated using both GPT-4-generated and human translations and were then asked to fill out a questionnaire evaluating five constructs: quality, comprehension, enjoyment, perception, and suggestions for improvement. The results indicate that GPT-4-generated translations can offer an equivalent or superior viewer experience to human translations. However, the results also highlight the potential of GPT-4-generated translation in humor translation and emphasize the need to refine AI algorithms, foster collaboration between AI and human translators, and establish quality standards for AI-generated translations based on audience preferences and evolving trends in humor.
Hussein Abu-Rayyash (Sun,) studied this question.