Artificial intelligence (AI) tools have quietly assumed a larger role in learning foreign languages over the past several years. As the quality of online machine translation (MT) has improved in recent years, how to use it for language learning purposes seems to be the “elephant in the classroom” (Loock, Lechauguette, & Holt, 2022), with a lack of explicit dialogue between teachers and learners regarding if, when, and how it should be used. There is evidence that students use online MT tools as a part of their language learning repertoire, but many teachers are still hesitant to allow them to be used officially (Ducar & Schocket, 2018; Lee, 2022). Similarly, an emerging new chat bot called ChatGPT was released in late 2022, causing an uproar in education with many institutions considering banning its usage over concerns that students may use it to cheat on exams and assignments. While research has started to identify areas where MT has the potential to be useful for L2 writing, particularly with lexical and syntactic aspects, as an emerging technology, the affordances of ChatGPT in L2 writing remain largely unexplored. It has the potential to encompass some elements associated with MT such as the ability to translate text on request, but it tends to do this at a more holistic level when compared to standard MT tools even including relevant citations and references according to designated formats. To raise language learners’ awareness of the pedagogical uses and potential pitfalls of AI tools, approximately seventy Japanese learners of English at a university in Japan were investigated for their current perspectives and practices with Deep-L and ChatGPT, as well as the impact of targeted training in their use. Strategies that focus on using these tools for planning, writing, and editing were provided. The study employs a mixed-methods approach, including: (1) learner attitude surveys; (2) analysis of writing tasks; and (3) observation of strategies used by learners. The processes and products of learners’ writing in English were analysed in terms of content, structure, and lexical and syntactic complexity and accuracy. The study seeks to explore the differences in attitudes towards academic English writing, MT, and ChatGPT, as well as examining both the writing processes and learners’ completed essays. Data collection is still ongoing, and preliminary results will be discussed in terms of the shifts in learner attitudes and behavior with Deep-L and ChatGPT as a result of the training.  

References 

Ducar, C., & Schocket, D.H. (2018). Machine translation and the L2 classroom: Pedagogical solutions for making peace with Google translate. Foreign Language Annals, 51(4), 779–795. https://doi.org/10.1111/flan.12366 

Lee, S.-M. (2022). L2 learners’ strategies for using machine translation as a personalized writing assisting tool. In J. Colpaert, & G. Stockwell (Eds.), Smart CALL: Personalization, Contextualization, & Socialization (pp. 184–206). London: Castledown Publishers.  https://doi.org/10.29140/9781914291012-9 

Loock, R., Lechauguette, S., & Holt, B. (2022). Dealing with the “elephant in the classroom”: Developing language students’ machine translation literacy. Australian Journal of Applied Linguistics, 5(3), 118–134. https://doi.org/10.29140/ajal.v5n3.53si2