Using screen recordings to assess how students use GenAI tools in the foreign language writing process : insights for language teachers
University students have overwhelmingly adopted AI tools when they write in their second language (L2). These include machine translation (MT) tools (Delorme Benites et al. 2021), which have been very popular among L2 learners for several years now, as well as new generative AI (GenAI) tools, such as general-purpose assistants like ChatGPT or Copilot as well as writing enhancement tools such as DeepL Write or Grammarly. Studies have shown that students are not necessarily GenAI literate (Barrett & Pack 2023), but not much is known about how they interact with these tools. Our presentation will report on data gathered in spring 2025 at a Swiss university to understand how students use different tools when writing in the L2. Using screen recording software, we recorded the writing process of 40 mostly French-speaking students when they produced a 300-word essay in English. The use of all external resources (GenAI and other language tools such as dictionaries) were coded based on a coding system for analysing information seeking activities (Hunziker Heeb 2020). We looked at the different reasons for using GenAI and how GenAI queries are related to the use of other external resources, such as MT or online corpora. In addition, the prompts produced by the students were compared and analysed for their complexity, such as their inclusion of contextual information or assigning a role to the GenAI tool. The results are merged to develop different writer profiles in the age of GenAI, akin to Hellmich’s (2021) development of MT user profiles. The results of this study reveal how students can use GenAI tools to improve their language production in ways that are conductive to learning. They also provide us with a better understanding of the type of training that students need in order to become literate users of GenAI tools, thus providing teachers with insights on how to incorporate these tools in writing tasks and courses.
Publication Details
- Journal
- Zürcher Hochschule für Angewandte Wissenschaften digital collection (Zurich University of Applied Sciences)
- Published
- 2025-11-14
- Primary Topic
- Text Readability and Simplification
- Type
- article
- Field-Weighted Citation Impact
- 0.00