University Students’ Use of ChatGPT: Insights and Prompt analysis : A multi-study on Students’ Opinions versus their Prompt Data

Sammanfattning

Generative artificial intelligence (AI) tools such as ChatGPT are increasingly present in higher education, yet limited research exists on how students use them in practice. This thesis investigates how, when and why students at KTH Royal Institute of Technology use generative AI in their studies. We employed a mixed-methods approach combining a student survey (n=96), a prompt submission form (n=93), and two focus group interviews. Quantitative data was analysed descriptively, while open responses and prompts were thematically categorised. The findings show that a majority of students use generative AI regularly. Most reported using it for conceptual understanding, coding support, writing assistance, and exam preparation. Thematic analysis identified eight key motivations, including speed, clarity, interactivity, and the ability to personalise explanations. 64% of respondents felt that AI helped them understand course material, and most who believed it affected their academic performance reported a positive impact. While concerns about overreliance and misinformation were noted, no evidence of academic misconduct was found. Interviews supported the survey findings and added further context on ethical reasoning and practical use. This study provides insight into current student use of generative AI and highlights the need for clearer institutional policies, pedagogical support, and continued research into its long-term effects in academic settings.