Meta-cognition on AI: What students think about using AI for academic purposes

There is currently a broad debate concerning the application of AI-based tools in academic contexts, particularly in the domain of academic writing of students, as well as concerning the AI-related skills necessary for these pursuits (cf. Long & Magerko, 2020). The utilization of Generative AI (genAI), underpinned by Large Language Models (LLM), holds considerable promise in facilitating academic processes, particularly for students to whom the language of academic study is not their first language (L1). A cross-sectional survey of students at German universities (von Garrel & Mayer, 2023) revealed that approximately two-thirds of respondents were already using genAI-based tools in the 2022/23 winter term, but only a quarter did so (very) frequently. Similar studies have primarily yielded a series of discrete snapshots of genAI utilization in academia. A notable limitation of these studies is the absence of any distinction between L1 and L2 students. The survey we conducted focuses on potential group differences between students with German as L1 and German as L2 and also aims to track developments in the use of tools based on genAI, knowledge about genAI, and attitudes toward genAI in the first years of the general availability of genAI-tools (2023-2025). To this end, a total of 143 questionnaires from students of various degree programs (primarily German as a Foreign and Second Language) from the University of Leipzig encompassing a period of two years are evaluated. The results of the survey presented in this paper concentrate on the students’ awareness of and disposition towards genAI / LLM (but see Ketzer-Nöltge & Rüger, in press, for complementary results).