<![CDATA[As AI translation tools increasingly enter translator training, understanding how students use them critically especially in literary translation has become essential. Literary texts require nuanced handling of culture, figurative language, and style, areas where raw AI output often fails. This qualitative descriptive study examined AI integration patterns among 18 undergraduate literary translators. With AI permitted, students completed five tasks: narrative analysis, element mapping, glossary building, translation, and written strategy justification. Open coding and textual triangulation revealed five patterns: (1) prioritizing communicative meaning, (2) domesticating while retaining cultural terms, (3) preserving religious-cultural elements, (4) documenting techniques item by item, and (5) producing strong outputs with minimal justification. A critical finding: translation quality alone cannot detect AI use; only written justifications make it visible. A three-tier developmental framework emerged: Exceptional (28%; 93–97/100), Strong (44%; 83–90/100), and Incomplete (28%; 37–65/100). Tier placement depended primarily on explicit grounding in translation theory (Newmark, Baker, Venuti). Theory-grounded students used AI more reflectively, treating outputs as provisional, recognizing untranslatability, and preserving cultural-spiritual concepts. Two phenomenological case studies (S1, S2) showed that translation excellence follows complementary pathways, with theoretical grounding as the common mechanism enabling critical AI use. The study recommends task designs that demand reflective transparency and pre-task theoretical instruction to foster critical AI literacy in literary translation pedagogy.]]>
Copyrights © 2026
