Department of Arabic Language and Literature

This thesis investigates the ideological shifts in human and AI-generated translations of Palestinian literature from Modern Standard Arabic into English and French. The corpus, which is based on two ideologically charged literary texts (Ra'aytu Ramallah by Mourid Barghouti and Tafṣīl Thānawī by Adania Shibli) encompasses the Arabic source texts, their human translations into English and French, and AI-generated translations produced using ChatGPT-4 Plus. The main goal of this research is to explore how trauma, displacement, and postcolonial narratives are mediated through both human and machine translation. This qualitative and comparative study is informed by Fairclough’s (1992) three-dimensional model of Critical Discourse Analysis (CDA), Baker’s (2006) ethics of re-narration, and postcolonial studies. The analysis is also guided by Calzada Pérez’s typology of ideological shifts to assess how meanings are constructed across languages. The preliminary findings reveal that while AI-generated translations are often syntactically and linguistically accurate, they tend to neutralize the ideological and emotional charge of the source texts. By contrast, human translations generally retain the ideological weight of the source text. The research also explores ethical considerations about the use of AI in translating ideologically and emotionally charged literature. The thesis contributes to the fields of translation studies by offering a nuanced analysis of how ideology and trauma are rendered or neutralized through different modes of translation, as well as the implications of using AI-generated translations for narrative agency in politically situated texts.