Opening the material session, Ms. Kiki outlined the real challenge teachers face: spending hours summarizing thick textbooks. Ironically, when seeking help from general AI, the answers often deviate or fabricate information due to its overly broad database sourced from the open internet.
As a solution, Ms. Kiki introduced the concept of Grounded AI through a free Google platform called NotebookLM. Unlike conventional AI, NotebookLM operates in a highly specific manner. This artificial intelligence only thinks, summarizes, and answers questions based on documents or learning resources uploaded directly by the user. This system drastically minimizes the risk of AI hallucinations or false data.
The online webinar atmosphere became more interactive during the hands-on practice session. Through a browser simulation, all participants were guided through data conditioning step by step:
1. Upload Module: Participants practiced uploading teaching material documents into the source panel. The AI immediately performed automatic summarization without disrupting the original material structure.
2. Discussion Based on Scientific Citations: Ms. Kiki demonstrated NotebookLM's main advantage, which is highly research-friendly. Every time a teacher submitted a command or question in the chat column, the AI's answer would display interactive citation numbers. When these numbers were clicked, the screen automatically scrolled to show the original paragraph from the source document. This feature greatly assists educators who wish to compose journal articles or validate teaching materials to maintain scientific accuracy.
3. Instant Conversion Studio: Participants experienced the greatest benefit when exploring the AI Studio feature on the right panel. In just minutes, teachers could produce a complete learning package. This package automatically includes a one-page summary, a list of key concepts, a glossary, and a ready-to-use multiple-choice quiz draft.