Implementing

I systematically integrated AI tools in developing my political science course materials. Below I outline the overall process used to create effective self-study resources with automated feedback.

Source Material Organisation

First, I compiled all my existing teaching materials that students used for their self-study. Most were in written form, but I also included video content and specific webpages to provide a comprehensive learning experience.

It proved essential to establish a clear file naming convention, as NotebookLM allows selective source utilisation. This organisation enabled the tool to indicate precisely which source contained specific information, facilitating quick reference attribution. For example, I labelled materials as “Democratic_Theory_Reading1.pdf” rather than generic names like “Reading1.pdf”, which helped both the AI tool and myself locate original sources efficiently.

Working with the Materials

NotebookLM effectively synthesised various sources to create questions that extended beyond mere fact-checking to require deeper conceptual understanding. I found it crucial to specify in my prompts which cognitive levels the questions should address (e.g., knowledge recall, analysis, synthesis, evaluation). For instance, when covering political systems, I explicitly requested:

“Generate questions that assess three levels of understanding: (1) basic knowledge of parliamentary systems, (2) analysis of similarities and differences between parliamentary and presidential systems, and (3) evaluation of which system might better serve emerging democracies.”

After determining the cognitive levels, I specified question formats (multiple choice, matching pairs, etc.). I then requested a specific number of questions for the self-study test, always including detailed explanations for incorrect answers. This detail proved particularly important, as without explicit instruction, the AI would often provide circular explanations (e.g., “This answer is incorrect because it is not right”).

Selection of questions and uploading in Moodle

The next step involved evaluating the generated questions, selecting the most effective ones, and requesting additional questions or refinements for specific topics as needed. Once I had compiled a sufficient question bank, I uploaded them to Moodle.

To streamline this process, I utilised our university’s question conversion tool, which requires questions in a specific format. I instructed NotebookLM to present questions in this exact format, enabling quick upload in XML format and making them readily available to students in Moodle.

Assistance for final exam

NotebookLM also proved valuable for summarising study materials by offering reflective questions that helped students prepare for summative assessments or final examinations. By instructing the tool to create an exam guide covering all provided materials, it produced comprehensive revision resources. I then used this guide to prepare balanced examinations, ensuring coverage of all course topics while maintaining appropriate distribution across the curriculum.

Through this systematic approach, I significantly reduced the time required to create high-quality self-assessment materials while improving the consistency and educational value of the feedback provided to my political science students.