Clash of AIs: Which translates Saudi children’s comics into English best?

Zemni, Bahia; Bounaas, Chaouki; Alshehri, Fadiah

Passerelle
Volume 14, Numéro 1, Pages 79-111
2025-12-12

Clash Of Ais: Which Translates Saudi Children’s Comics Into English Best?

Authors : Zemni Bahia . Bounaas Chaouki . Alshehri Fadiah .

Abstract

This study investigates the performance of Large Language Models (LLMs)—ChatGPT, DeepSeek, and Claude—in translating literary texts, with a particular emphasis on children’s literature. Recognizing the unique challenges posed by this genre, such as age-appropriate content, expressive language, and cultural nuances, the study employs Juliane House’s Translation Quality Assessment (TQA) model to assess the quality of AI-generated outputs. Content and comparative analyses will be conducted on selected examples from Saudi children’s comics translated by the three tools. The outputs will be evaluated based on linguistic accuracy, stylistic fidelity, and contextual appropriateness. The findings highlight strengths and limitations in each tool's performance in handling the intricate demands of literary texts, providing practical insights for translators and students regarding tool selection and effective usage.