Limitations of Machine Translation of Egyptian Dialect in Social Comedy Series into English: A Comparative Study between AI Subtitling Tools.

Document Type : Original Article

Author

Translator)

Abstract

This comparative study investigates the limitations of Machine Translation (MT) in subtitling Egyptian Arabic comedy series, with a particular focus on the preservation of pragmatic meaning and cultural humor. The selected dataset extracted from Egyptian social comedy shows, subtitled using three AI tools: FreeSubtitles.ai, Veed.io, and Maestra.ai. These machine-generated subtitles are evaluated against the theoretical framework of Mona Baker’s (1992) pragmatic equivalence model, an extended version of Pedersen’s (2017) FAR model and Gottlieb’s (1992) subtitling strategies. Through a qualitative and quantitative comparative approach, the AI subtitles are contrasted with human translation strategies to assess how well meaning, tone, and cultural references are conveyed. The findings indicate that MT tools frequently rely on literal translation, often failing to capture the implicit humor and cultural nuances embedded in Egyptian colloquial Arabic. This study highlights the need to expand AI training datasets to include dialectal Arabic, not just Modern Standard Arabic to improve the quality and contextual accuracy of subtitle translations. Ultimately, the research affirms that human translators remain essential in delivering contextually appropriate and culturally sensitive subtitles in comedic content.

Keywords

Files

Share

How to cite

Statistics