In an era where international education trends increasingly prioritize the integration of artificial intelligence (AI), there is a critical need to understand how students effectively use these tools to foster innovative and deep learning. This study addresses this gap by investigating higher education students’ experiences with advanced AI tools within a nine-week instructional experiment structured by the Predict-Observe-Explain (POE) model. Our primary motivation was to explore how a structured, inquiry-based framework could scaffold the development of sophisticated AI literacy, guiding students toward strategic human-AI partnerships. We employed a qualitative case study design, collecting data from 17 students at a Taiwanese university through written focus group interviews. Participants were granted free access to premium-tier generative AI tools, including ChatGPT and NotebookLM. Findings reveal that students developed sophisticated, task-aligned workflows by strategically combining multiple AI tools, a progression significantly accelerated by institutional scaffolding. Participants reported substantial benefits, including enhanced efficiency and deeper cognitive engagement, while also navigating persistent challenges such as accuracy concerns and technical limitations. They adopted adaptive strategies, including cross-tool verification, prompt refinement, and critical evaluation, to mitigate these issues. The study further demonstrates that AI tools were particularly effective in supporting research question refinement and academic reasoning. This research makes several contributions to the field of educational technology. It provides empirical evidence that inquiry-based models like POE are effective for guiding AI tool integration and fostering higher-order cognitive skills.
- Inicie sesión para enviar comentarios