Abstract This paper assesses AI-assisted translation by viewing prompting as an externalisation of cognitive processes during translator-AI interaction. More than mere technical inputs, prompts serve as authentic written traces behind the identification of problems by translators and their resultant solutions. This approach differs from traditional think-aloud methods by offering explicit access to decision-making. Using screen recordings and think-aloud protocols, the study observes seven trainee translators working on English-to-Chinese translation tasks using ChatGPT. The paper analyses three critical decision points: (1) when outputs are accepted, (2) factors triggering prompt engineering, and (3) criteria leading translators to stop prompting. Findings provide evidence-based insights for developing translator AI competence and inform a pedagogical framework that highlights the visibility of prompting processes, thereby enhancing translator education in the AI era.
Shuyin Zhang (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: