The integration of Artificial Intelligence (AI) into the scientific writing process presents both a significant challenge and a didactic opportunity for higher education (see eg. Abd-Elsalam & Abdel-Momen, 2023; Farhat & Arafa, 2024). At the Zurich University of Applied Sciences (ZHAW), a design-based research approach was employed to develop a workshop-based, student-centered course that facilitates the reflective use of AI tools in various phases of scientific writing for engineering students. The initial course implementation in spring term 2024 focused on systematically incorporating AI into the writing process, equipping students with both methodological and technological competencies (Schnatz & Bieri, 2024).

Building on the findings of this first iteration, the course design was further refined for the second implementation in the spring term of 2025. A prompting journal was introduced, requiring students to systematically document their use of AI throughout the writing process. This journal serves as both a reflective tool and a means of gathering insights into how AI influences the structural aspects of scientific writing, such as the introduction and discussion sections. By analyzing these journals, we aim to explore correlations between AI usage patterns and improvements in the coherence, clarity, and organization of students’ writing. Additionally, self-learning units on the central chapters of a scientific paper were added, in which students, for example, work on the CaRS model for the introduction via videos, texts, and interactive tasks. In a joint project between the ZHAW and the Zurich University of Teacher Education (PHZH), an AI tool was developed specifically to support students’ introduction writing based on the CaRS model (INTR-O – Introduction Optimizer, Businger et al.)

This paper presents insights from the ongoing second iteration of the course, focusing on how structured AI interventions - such as the prompting journal and targeted AI tools - shape students’ engagement with scientific writing. By examining students’ reflective documentation and learning progress, we seek to understand how AI-assisted approaches contribute to writing development. These findings will contribute to the evolving discourse on AI in academic writing education and inform future iterations of the course, refining the balance between AI-supported and conventional writing strategies.

Keywords: Scientific writing, AI integration, AI literacy, student-centered learning, design-based research, course design, engineering education.