Yangsi Liao
Thesis title (PhD): Investigating the Influence of AI-generated Captions on Interpreters’ Cognitive Load and Performance in Onsite Simultaneous Interpreting: A Case Study on English–Chinese Culture-Specific Items
The rapid development of artificial intelligence has led to the increasing use of AI-generated captions in simultaneous interpreting, alongside slides and speaker videos. Source-and target-language captions may support interpreters’ comprehension, anticipation, and interpreting performance. However, they may also increase task complexity, as interpreters must coordinate auditory input, multiple visual sources, and target-language production under time pressure.
These challenges may be particularly salient when English culture-specific items (CSIs) are interpreted into Chinese. Such items often carry culture-bound meanings and may require interpreters to select an appropriate rendering strategy. Despite growing interest in AI-assisted interpreting, limited empirical research has examined how caption language shapes interpreters’ attention and performance when culturally specific information is involved.
This study investigates the effects of AI-generated captions on interpreters’ visual attention and rendition of CSIs during onsite English-to-Chinese simultaneous interpreting of cultural diplomacy discourse. It addresses three research questions:
RQ1: How does the availability of source- and target-language captions affect interpreters’ visual attention allocation among captions, slides, and the speaker when culture-specific items (CSIs) occur in English–Chinese simultaneous interpreting of cultural diplomacy discourse?
RQ2: How do source- and target-language captions affect the completeness of interpreters’ rendition of culture-specific items (CSIs) in English–Chinese simultaneous interpreting?
RQ3: To what extent are interpreters’ gaze behaviours towards captions, slides, and the speaker associated with the quality of their rendition of culture-specific items (CSIs), and do these gaze–output relationships differ between source- and target-language caption conditions?
Approximately 15–20 interpreters will undertake English-to-Chinese simultaneous interpreting tasks with source- and target-language captions. Eye-tracking data will examine attention allocation across captions, slides, and the speaker, while interpreting outputs will be analysed for CSI rendition quality and completeness. Post-task interviews will explore participants’ experiences of caption use. The findings will contribute to research on AI-assisted simultaneous interpreting and inform interpreter training and user-centred caption design.
Supervisory team
- Dr Jade (Biyu) Du, School of Modern Languages, Newcastle University
- Dr Yalta (Yayun) Chen, School of Modern Languages, Newcastle University