Disambiguating Ambiguity in Academic Writing Using AI: Insights from German and English Language Students at the Faculty of Philology, Goce Delcev University, Stip, North Macedonia

This study investigates the application of Artificial Intelligence (AI) in identifying and resolving lexical and syntactic ambiguity in academic writing, with a focus on German- and English-language texts. Ambiguity poses a significant challenge to clarity and precision and may lead to misinterpretation in scholarly communication. The research examines how AI-based language models, particularly ChatGPT, can support students in detecting ambiguous expressions, evaluating alternative interpretations and reformulating them with greater precision. The study involved 30 students from the Faculty of Philology, Stip, including majors in German Language and Literature, English Language and Literature, and students studying German or English as elective foreign languages. Participants completed four task types involving ambiguity detection, word-sense disambiguation, translation of ambiguous sentences, and critical evaluation of AI-generated translations. Their responses were compared with AI-generated analyses and reformulations to examine the respective contributions of human judgment and AI-assisted language processing. The findings show that students successfully recognized most ambiguities when sufficient contextual information was available, while AI proved particularly effective in generating a broader range of possible meanings and interpretations. In contrast, students outperformed AI in tasks requiring contextual and pragmatic reasoning, particularly when evaluating and correcting AI-generated translations. The results suggest that AI is a valuable pedagogical tool for supporting ambiguity awareness and academic writing, while human contextual interpretation remains essential for accurate scholarly communication.