Comparing learners’ engagement strategies with feedback from a Generative AI chatbot and peers in an interpreter training programme: a quasi-experimental study

Interpreter training programmes worldwide commonly adopt a learner-centred approach, incorporating peer feedback and technology-assisted feedback. However, the potential of Generative AI in providing two types of feedback, evaluative comments and learning strategies has received limited attention. This quasi-experimental study fills this gap by investigating and comparing interpreting students’ engagement strategies with feedback from a Generative AI chatbot (n = 46) and their peers (n = 42). Participants were encouraged to seek feedback based on the transcripts of their recorded interpretation performances. Using a pre-designed reflective sheet, their engagement strategies with feedback were quantitatively analysed. After controlling for feedback literacy level and academic performance, MANCOVA tests suggested that, compared with the peer feedback group, interpreting trainees interacting with the Generative AI chatbot reported a significantly lower level of agreement and a higher level of dispute with the received evaluative comments. Participants were more motivated to argue with the chatbot than with their peers. Furthermore, learning strategies provided by the chatbot were more likely to be both accepted and disregarded by participants than those from peers. The potential of using Generative AI chatbots as a supplementary source to support peer and technology-assisted feedback practices in interpreting programmes is discussed.