Supporting Authorship and Reflection in AI-Assisted Academic Writing

AI writing tools are becoming increasingly integrated in university students’ academic workflows. While these tools support productivity, they often reduce students’ sense of authorship and engagement in the writing process. This thesis examines how interaction design can support student authorship and reflective use of AI in academic writing. The study follows the Double Diamond methodology and includes qualitative methods such as interviews, writing observations, tool analysis, a co-creation workshop, prototyping, and usability testing. The research was conducted with students from Malmö University. Based on the findings, a mid-fidelity prototype was developed and tested, focusing on features such as visible AI-generated content, rewrite in your own voice functionality, and AI-powered peer review with explanations. The results suggest that these features can help students reflect on and maintain control over their writing when using AI. The study highlights the importance of designing AI tools that support learning and transparency rather than automation alone.