Integrating Generative AI in Higher Education: Practical Applications and Institutional Guidelines

The rapid advancement of Generative AI (GenAI) presents both opportunities and challenges for higher education. Traditional educational methods often struggle with processing large datasets and facilitating real-time interactions, areas where AI offers potential solutions. This paper examines the development, application, and evaluation of a custom GPTs-based AI teaching assistant, the DS-ASST app, specifically designed for Data Science education within a university liberal arts context. Leveraging Retrieval-Augmented Generation (RAG) technology integrated with course-specific materials, the system aims to enhance both instructor efficiency and student learning experience. The study evaluates the AI assistant's impact across four key areas: teaching preparation efficiency, active learning support, data analysis process enhancement, and the promotion of advanced learning activities. Findings indicate significant improvements in instructor workflow, allowing for more focus on pedagogical refinement rather than routine content creation. For students, the tool provided on-demand concept clarification, guided problem-solving, and personalized learning path suggestions, fostering self-directed learning and engagement, particularly with complex data analysis tasks. The system also aided in developing practical data analysis skills through workflow guidance and interpretation support. Technical challenges inherent in using Large Language Models (LLMs), such as optimizing prompt design for educational relevance and mitigating the risk of AI "hallucinations," were addressed through systematic experimentation and the integration of RAG with a verified knowledge base. Furthermore, the paper discusses the broader implications of GenAI in education, including the redefinition of instructor roles and the evolution of assessment methods. At the same time, based on the implementation experience and existing frameworks, the study proposes practical institutional guidelines and checklists for the ethical and effective use of GenAI by both faculty and students in university settings, emphasizing academic integrity, critical thinking, and AI literacy.