LLM usage in Education

This thesis explores the potential risks and implications of deploying generative AI and opaque models within educational settings. It delves into the inner mechanisms of generative AI to uncover the complex processes that drive these advanced systems. The research highlights the unintended negative impacts these technologies can have on decision-making processes. Through a comprehensive analysis, this thesis proposes viable solutions and best practices to mitigate these risks, ensuring the responsible adoption of AI in schools. Furthermore, it investigates strategies to enhance the effectiveness and utility of responses from large language models. This study aims to provide a framework for students to harness the benefits of AI while safeguarding against its pitfalls.