Generative AI in Student Project Reports

Preliminary insights on Patterns of Use and Documentation in Higher Education

Abstract

Generative artificial intelligence (GenAI) tools such as ChatGPT challenge how knowledge, authorship, and learning are declared in higher education (HE). Written projects have long functioned as the primary medium through which students demonstrate independence, originality, and academic judgment. But as GenAI is integrated into teaching and learning, HE is transformed. This paper examines how students at Aalborg University declare their use of GenAI in project reports and what these documentation practices reveal about the digital transformation of HE. The analysis is based on a dataset of 5.222 publicly available student projects submitted between January 2024 and March 2025. To do this, the study employs a Python-based text analysis pipeline to identify and extract mentions of GenAI in student projects. A total of 787 projects (15.1%) explicitly mention GenAI tools. Projects were divided into low-frequency (1–20 mentions) and high-frequency (21+ mentions) groups, with the low-frequency group (n = 670) forming the basis for a qualitative thematic analysis of 1,152 mentions of GenAI in students’ projects. Analytical categories were developed inductively following the principles of reflexive thematic analysis supported by content analysis. Methodologically, the study demonstrates the potential of large-scale text extraction to examine educational transformation as it materializes in academic writing. The findings show that students use GenAI across a range of activities ranging from surface-level writing support to analytical and methodological applications. Particular attention is given to the theme ‘Going Beyond Writing with GenAI’, where students employ GenAI to refine coding frameworks, structure research designs, and critically reflect on methodological choices. These practices suggest that GenAI is beginning to take up methodological space in project work, functioning as a co-constructive partner in analysis and design rather than just a writing tool. Students also display awareness of GenAI’s limitations, such as its opacity and the “black box” problem, indicating an emerging form of critical AI literacy grounded in reflexivity and transparency. While the method captures only what students choose to declare, this unveiling of students’ usage is valuable, as well as the revealing of what students themselves consider worth declaring in their engagement with GenAI. Findings from the study indicate that the documentation regarding GenAI use is an epistemic and communicative practice through which students negotiate technological agency, authorship, and learning in a digitally transforming landscape of HE.