AI-augmented learning using a Retrieval Augmented Generation (RAG) Large Language Model (LLM)

This article is part of a series of case studies from McMaster instructors that explore perspectives and approaches for integrating or limiting the use of AI in learning environments.
–

Dr. Mike Welland from the Engineering Physics Program in the Faculty of Engineering has utilized Google’s Gemini Notebook (formerly NotebookLM) to augment his Numerical Methods for Engineering (ENGPHYS 3NM4/2NM4) course with GenAI tools. As a Retrieval Augmented Generation (RAG) AI, Gemini Notebook can utilize a range of sources selected by the instructor to present an AI chatbot for students and can even generate its own educational content. The generated responses and content are linked with the source material provided with reliable references to the source.

The New Process

Administrative Setup Process:

• Dr. Welland found the administrative setup process to be very simple
• The instructor needs to register for a Google Account to create the notebook with a limited number of sources.
  • A paid version of the Gemini Notebook is required to increase the number of sources attachable and to be able to share a link to the module with the students. Gemini Notebook membership
  • Students can access the completed notebook for free with a separate Google account.
• Content sources are added to the notebook and can be selected/de-selected as desired by the creator. E.g.: course notes, textbooks, youtube videos (including recorded lectures), podcasts, websites, etc. Video / Audio sources are automatically transcribed when added with timestamps facilitating backward referencing.
• The rest of the process is automated within the notebook
• The students are provided with a hyperlink to the notebook or could make their own with the source material

What does the notebook offer?

• An AI chatbot (based on Google Gemini) is created that uses the provided sources to build structured overviews of course content and answer student queries
  • Gemini Notebook uses Retrieval Augmented Generation (RAG) and the Google Gemini base for its tailored responses
• The AI chatbot will provide reliable references to the specific sections of the sources it used to answer a query
• Additionally, the tool can create AI-generated content for students to use:
  • Full slideshows, flashcards, and video overviews covering core concepts
  • Podcasts with two hosts discussing the engineering concepts. Students are able to ‘call-in’ to speak with the podcast hosts directly
  • Interactive mindmap summarizing the structure of the course and core concepts
  • Unlimited practice exam questions and multiple-choice quizzes can be generated for students to practice their understanding of the topic and prepare for exams

Test the notebook for yourself

• Try the Numerical Methods for Engineering (ENGPHYS 3NM4) AI-augmented notebook
• See Dr. Welland’s video on AI in Engineering Education for additional examples of his work

The Impact and Feedback

Students had access to the tool throughout the length of the course but tracking data showed that students primarily used the AI tool in the 2 weeks leading up to the final exam with over 200 unique queries made of the AI chatbot, approximately 50% of these in the days immediately preceding the exam.

When surveyed about the effectiveness of the Gemini Notebook module, students indicated that they had primarily used it to understand specific sections of the course content and to utilize the practice exams and multiple-choice quizzes. This supported the students in learning the course content, with one student indicating that they were able to understand the course content better because the repetition of the practice assessments had helped cement some of the core engineering concepts.

However, students found the visual and audio content generated by the AI system (such as the podcasts and video tutorials) to be unappealing. The students found the robotic nature of the delivery to be less engaging than existing high-quality human-generated YouTube tutorials.

Ongoing Work

Dr. Welland has indicated that he would be happy to use the Gemini Notebook system in the future since it is easy to set up and has provided some positive learning results for his students. He views this as the first step towards Intelligent Tutoring Systems (ITS) that can provide tailored education to each individual student based on their own learning methods and preferences.

In the future, Dr. Welland is hoping to collaborate with other engineering instructors to bring similar RAG AI systems into large engineering classes to provide several hundred students with AI-augmented learning opportunities. There are also many opportunities for application to general information access systems (e.g.: academic calendars, policy documents, etc).

—

Welland, M. (2026), AI-augmented learning using a Retrieval Augmented Generation (RAG) Large Language Model (LLM). Licensed under Creative Commons BY-NC-SA 4.0.