How AI Translation Personas Reframe Diplomatic Texts

Abstract

This study explores how persona prompting in AI translation systems, particularly OpenAI’s ChatGPT (GPT-4o), shapes politically sensitive discourse. By assigning translation personas such as “diplomat” and “politician,” the AI produced distinct Arabic renderings of six controversial statements by Prince Zeid Ra’ad Al Hussein, former UN High Commissioner for Human Rights. Drawing on Descriptive Translation Studies, Critical Discourse Analysis, and Narrative Theory, the analysis examines tone, word choice, and rhetorical strategies. Findings show that AI translations not only transfer meaning but also reframe ideas: diplomats soften charged language, while politicians balance criticism and formality. These shifts highlight translation as a political and ethical act with implications for global perceptions, policy, and public opinion. The opacity of AI processes further raises concerns over transparency, accountability, and bias. The study contributes to research at the intersection of computational linguistics, translation studies, and political discourse, urging critical scrutiny of how AI translation shapes meaning and ideology in international communication.