GENERATIVE AI FOR WRITTEN ASSIGNMENTS IN HIGHER STEM EDUCATION: EXPLORING THE INTERACTION BETWEEN STUDENTS AND AI THROUGH CASE STUDIES

The application of Large Language Models (LLMs) through generative AI tools in higher STEM education is reshaping academic writing, offering both opportunities and challenges. This study explores how students use publicly available generative AI tools, such as ChatGPT and Microsoft Copilot, for written assignments through two case studies.

We examined students’ intents, engagement strategies, and reliance on AI-generated feedback using survey data and chat history analysis. Our findings reveal distinct usage patterns, from text assistance to problem-solving and factual inquiry, highlighting both the benefits and limitations of AI as a learning tool.

Additionally, interviews with teachers provided insight into the instructional design considerations for allowing generative AI tools during coursework. We aim to inform best practices for balancing AI support with skill development by examining when and how these tools should be incorporated. Understanding these dynamics will help educators create effective AI-assisted learning environments while addressing concerns about overreliance and academic integrity.