Generative Artificial Intelligence (GAI), powered by Large Language Models (LLMs), has rapidly entered higher education, offering instant content generation but raising concerns about student overreliance and cognitive slacking. While GAI outputs often appear credible, they remain vulnerable to bias and hallucinations, creating risks of misinformation. Current teaching practices focus mainly on tool operation, lacking systematic approaches to guide students in verifying authenticity or engaging in critical evaluation. This study proposes the Multi-model Deliberative Generative AI Framework (MDGAIF), which employs cross-model verification and structured guidance to transform students from passive recipients into active evaluators and ethical stewards of AI. A quasi-experimental design was conducted with students in an “Introduction to Machine Learning” course, comparing an experimental group (MDGAIF-integrated teaching) with a control group (traditional tool-based teaching). Mixed-methods analysis revealed that MDGAIF significantly improved students’ habits of verifying information and enhanced their critical thinking and problem-solving skills. Findings demonstrate that MDGAIF not only achieves its intended goals but also provides a valuable reference for universities seeking to design integrated AI teaching strategies and curricula.

Authors:
Shyh-Jian Tang, WuFen University, Taiwan

About the Presenter(s)
Tang Shih-Chien,Associate Professor of Professional Practice in the Department of Artificial Intelligence and Drone Technology at Wufeng University. I have integrated various GAI platforms and tools into my curriculum.

See this presentation on the full schedule – Sunday Schedule

Bookmark