Abstract:

Teaching staff at universities is obliged by a recent law to get trained in skills related to the use of Artificial Intelligence in the context of teaching and university-related activities. At the same time, students also have to be trained in using AI for the improvement of their academic work, try the potential and limits of the technology, and reflect on its use. A recent paper (Leiss 2024) examined AI's potential in improving students’ reflection and progress in self-determined goal setting. Building on these insights, the present contribution explores AI’s role in fostering academic writing and scientific thinking.

One of the recent attempts to improve the didactics of scientific writing is to start in the early semesters with the development of essay skills and small academic writing tasks of different kinds. The concept emphasizes the importance of enhancing variability and individuality in students' tasks to increase motivation and quality, by providing a substantial amount of autonomy and self-determination. It is assumed that several little tasks (resulting in a portfolio to be graded as a whole), preferentially connected, can help students learn and develop scientific skills at the same time. The proposed contribution reports on ideas for small tasks enhanced with AI that not only enable students to grow in terms of knowledge but also in terms of academic writing skills.

A total of 29 first-year students were instructed to develop three research questions on the topic of motivation, drawing on motivational theories introduced in class. Additionally, they were asked to apply one of the questions to themselves. The assignment was to be presented in minimum three PowerPoint slides. Students were required to use ChatGPT via the university’s website as a supportive tool. Importantly, all input was expected to originate from the students themselves, with AI serving only as an auxiliary resource. Each student was asked to document three successful prompts and respond to evaluation questions regarding their experience.

Findings indicate that students were generally enthusiastic about using AI tools; however, they encountered difficulties in formulating effective prompts. Nearly half of the participants did not provide any prompts at all, and weaknesses were evident in many of the prompts that were submitted. Moreover, it became clear that AI was often employed less as a supportive instrument and more as the primary author of the text. This finding is backed up by evidence that students overwhelmingly did not include their knowledge so far on motivation theories in the prompts they used for ChatGPT.

The implications suggest that integrating AI into coursework can foster digital literacy and reflective learning, but also highlights the need for explicit training in prompt design to ensure AI is used as a supportive tool rather than a text generator.

Keywords:

Generative AI, Pedagogical Innovations.