Human-centered pedagogies in the age of generative AI

Abstract Since its release in November 2022, ChatGPT has been used by students alongside other AI-powered tools like machine translation for various language-related tasks. Despite its growing use, educators and researchers have not yet fully monitored or understood its impact within academic settings. This study investigates the perceptions of undergraduate students enrolled in a translation course in an English Degree program, focusing on how they view, experience, and use generative AI, with particular emphasis on ChatGPT as the most widely used tool. Conducted as part of the teaching innovation project Multilingual Competence: Implementing AI (ChatGPT) for Multilingual Classroom Success at Universitat Rovira i Virgili, Spain, this study spanned five 90-minute sessions, during which students engaged with ChatGPT through three types of exercises that involved translation, post-editing, and comprehension of texts generated by this tool. Pre- and post-experiment questionnaires were administered to examine the impact of ChatGPT on students’ perceptions of progress and sense of agency. The findings indicate that students generally hold neutral-to-positive views regarding the effectiveness of ChatGPT in translating, writing texts, and language learning. However, ChatGPT received particular criticisms as a post-editing tool, as evidenced by quantitative as well as qualitative data. The students emphasized the need for additional training, particularly in prompt generation. On the other hand, some students expressed their concerns regarding data privacy or various ethical issues such as the environmental impact of ChatGPT. In terms of agency, quantitative and qualitative data shows that most students believe that they retain significant control in their interactions with ChatGPT. These results suggest that while students recognize the potential of ChatGPT, they are also aware of its limitations. These findings align with Human-Centered Artificial Intelligence (HCAI) approaches, which emphasize the importance of human control and critical thinking as fundamental principles in fostering effective and responsible human-machine interactions.