Abstract Book of the 9th International Conference on Teaching, Learning and Education
Year: 2025
[PDF]
AI-Supported Multilingual Explanations: Investigating ChatGPT’s Role in Multilingual Interactions
Kuzu, Taha Ertuğrul (Jun.-Prof. Dr.)
ABSTRACT:
Artificial intelligence (AI) holds significant potential in language learning as well as multilingual contexts due to its capacity to translate fluidly. However, a usage in pedagogical contexts requires the development and exploration of adequate learning-environments if AI-systems are intended to be used in school contexts for scaffolding multilingual (trans-)languaging processes. With the recent ability to translate both spoken and written language, various didactic scenarios for supporting multilingual languaging processes are possible: E.g., learners can use AI to collect necessary terms for multilingual explanations, or they can get reflection impulses regarding nuanced meaning differences between languages according to the linguistic relativity hypothesis. However, explorative insights are necessary into the conditions as well as chances and hurdles of such an implementation of AI into learning environments. In this study from Germany, such an implementation is evaluated: ChatGPT was used by nine Primary school learners to generate multilingual storyboards for explanatory videos about the mathematical compensation strategy. A Design-based research study was conducted by iteratively developing and evaluating a learning environment that includes the usage of ChatGPT. For the analysis of the processes, interpretative interaction analyses were conducted to gain abductive insights into the usage of AI. The main findings are that the usage of ChatGPT appears particularly beneficial at the end of teaching units to ensure that the learners have pre-knowledge when prompting and that a prior didactically reduced prompt engineering training is necessary for ‘fruitful’ interactions with AI, as the comparison between trained and untrained groups indicates.
Keywords: AI in education, translanguaging, design-based research, interpretive analyses, prompt engineering