Populist political discourse poses a particular challenge for translation because its persuasive force depends not only on propositional content but also on evaluative language, emotional appeal, and the construction of antagonistic identities. As Large Language Models (LLMs) become increasingly integrated into translation workflows, questions arise regarding their capacity to reproduce or reshape the ideological and stylistic features of politically charged discourse. While prior research has focused primarily on translation quality and lexical accuracy, less attention has been paid to the role of prompt engineering in shaping ideological framing in LLM-generated translations.
This study investigates the translation of Greek populist political discourse into English and German using the Gemini LLM. It examines whether altering the prompt-induced translator persona, ranging from a neutral policy voice to a politically engaged populist stance, systematically affects the stylistic, emotional, and ideological features of translated texts. It further explores whether such effects differ across target languages and across distinct forms of populism.
The corpus consists of nine political texts by three Greek politicians representing exclusionary, institutional, and punitive populism. Each text is translated under three prompt conditions, namely a neutral policy translator, a concerned citizen, and a populist firebrand. A mixed-methods design combines quantitative and qualitative approaches. Quantitative analysis measures lexical, syntactic, semantic, and sentiment variation using established evaluation metrics, specifically BLEU, TER, COMET, and BLEURT, alongside corpus-based tools such as Sketch Engine and NRCLex. Qualitative analysis employs Critical Discourse Analysis and framing theory to examine shifts in evaluative language and political representation.
Findings contribute to debates on translator agency, algorithmic mediation, and ideological bias in AI-assisted translation, demonstrating that prompt engineering can function as a mechanism of ideological modulation in LLM-based translation systems. More broadly, the study highlights prompt engineering as a potent form of ideological intervention within contemporary translation practices.
Michail Georgios Dimelis is an undergraduate student in the Department of Foreign Languages, Translation and Interpreting at Ionian University, specializing in economic, legal, and political translation (Greek, English, and German). To further enrich his linguistic and technical expertise, he completed a six-month Erasmus mobility program in Saarbrücken, Germany. His research interests lie at the intersection of digital humanities and translation, specifically exploring the application of AI and its impact on the translation of politically charged texts. Alongside his studies, Michail serves as the student coordinator for the volunteer initiative “DfltiVoice”, organizing webinars and activities that highlight the modern relevance of translation. To bridge the gap between academic theory and industry practice, he recently participated in the Mentoring Program of the Panhellenic Association of Professional Translators Graduates of Ionian University (PEEMPIP). Additionally, he served as the EU Student Ambassador at Ionian University for the 2025-2026 academic year, promoting career opportunities and fostering a deeper understanding of EU democratic values.
Stavroula Paraskevi Vraila is an Assistant Professor in the Department of Foreign Languages, Translation and Interpreting at the Ionian University, specializing in Economic, Legal and Political Translation (Greek–German). She holds a PhD from the National and Kapodistrian University of Athens and a Master’s in Translation Studies from the Ionian University. Her research focuses on the translation of political and legal texts, and the applications of AI in translation, exploring issues of political ethics and digital bias. She has published widely in Greek and international journals and presented at global conferences. At the university, she teaches undergraduate and postgraduate courses and actively participates in departmental committees (Erasmus, Evaluation, Outreach). Since 2007, she has also worked as a professional translator with publishing houses and digital companies, specializing in institutional and communicative texts with an emphasis on accuracy and functional effectiveness.
Aristidis (Aris) Vrahatis is an Assistant Professor of Artificial Intelligence in Modeling Complex Systems, in the Department of Informatics, Ionian University, Corfu, Greece, and a member of the Bioinformatics and Human Electrophysiology Laboratory in the same Department. He received his Ph.D. in 2016 from the Department of Computer Science and Engineering, University of Patras, Greece, where he graduated in 2011 and received his M.Sc. Diploma in the program ‘Computer Science and Technology’ in 2012. He has more than 40 publications in journals and conference proceedings, including Oxford Bioinformatics, BMC genomics, IEEE journal of biomedical and health informatics, and Cellular and molecular life sciences. He has worked for more than 10 years in research and development projects, funded by the EU and other bodies. Research Interests: Machine Learning, Biomedical Data Mining, Biological Complex Systems Modeling, Network Medicine
