Prompts and Play: Teaching Worldbuilding and Fandom with NotebookLM

ChatGPT said: Here’s an alt text description for the image: "Two individuals, one standing and one sitting, interact with futuristic holographic screens showing various vibrant, digital worlds. The image is divided into a tiled layout of nine screens, each depicting a different scene—ranging from futuristic cities with glowing lights and towering structures to lush, green landscapes. The individuals are engaged in worldbuilding, surrounded by vivid, neon-colored environments that hint at advanced technology and creative exploration.

As instructors grapple with the fast-paced development of artificial intelligence (hereafter referred to as AI), one of the concerns that gets repeated is the question of critical attention to content. With all of the different programs out there, how do we ensure that students are reading and engaging with the material? And as there seems to be a new AI every day, how do we keep up with this pace of change?

One of the newer players on the field, Google’s NotebookLM, offers extensive possibilities. NotebookLM works on a “folder” based system, and allows you to curate your documents, websites, and other content and generate content based on the materials you provide. As you add documents, users have several ways to engage, including study guides, discussion questions, and lecture notes. What makes NotebookLM uniquely exciting, however, is the podcast feature. Once you’ve added your documents, you can generate a podcast based on your research documents. It’s a “deep dive” into the topic you’ve prompted, and sounds like any other podcast you can find on Spotify or Apple.

It is this feature that holds the most potential… and the feature that promotes new questions and anxieties for instructors. If they have a podcast, is this just a new way of enabling students to skip reading? Will students use this tool as a new workaround to course assignments? While these are reasons we should be cautious, all of these are also questions we should confront because there is incredible potential for creating collaborative, accessible research in class.

In this article, I’ll describe how I teach using NotebookLM extensively in two separate courses, and I will discuss how engaging with content in notebooks helps students both learn and create. First, I will describe how students learn about and engage in worldbuilding for my Games and Culture course. Then, I will describe how it supports a semester-long research project in my Fan Studies course. Both of these are entry-level courses within a Communication and Media Studies program. Students learn about design, history, and fandoms as they discuss and discover theories about media and communication. For both courses, the assignments I will discuss represent significant challenges as students are asked to be integrate and engage with complex and open ended ideas. Feedback in previous courses shows that students struggle to see the connections and synthesize ideas as they move through the course sequence. I argue that NotebookLM offers a way to address these issues, and offers an opportunity to bring collaborative research into the classroom.

How it works

NotebookLM is based on a folder system called, as you probably guessed, notebooks. For any new topic, you first create a notebook and then upload, link, and write to provide useful material about the topic. The AI is designed to work within the notebooks themselves, using the material you share, to craft its outputs. Once you have your sources, websites, etc. added to the notebook, you can begin to use the generative AI features.

Fig. 1 NotebookLM dashboard page
Fig. 2 NotebookLM upload screen

The main screen includes a list of your sources to the left, a chat bot in the middle, and the studio on the right. Google gives you a few standard options to try in terms of generating documents, including the study guide, FAQ, and briefing document. The chat bot allows you to input your own prompts as well. The studio is the running record of the documents you generate, and also includes the audio overview feature. The audio overview scans the documents you’ve uploaded and selected, and creates an incredibly realistic podcast that covers the main concepts in the text(s) in layman’s terms.

Fig. 3 Audio overview

World-building as Course Assignment

In my Games and Culture course, the worldbuilding assignment is the first major project. It asks students to create a fictional world for their game to take place in, and address aspects such as societal structures, geography, economy, government, etc. Students are assessed on the content, as well as the complexity, of their world. In past courses, one of the obstacles for many students is achieving a cohesive, consistent world that integrates new factors that make their world unique.

Generally, students work on this project for four to five weeks during the latter half of the semester, with the deliverable being a website and presentation. Developing a concept for a video game, like developing the concept for a novel, a movie, or other complex work, often starts with a fairly interesting but basic idea. The challenge of this assignment is to fill out the idea, to situate their characters and themes in an engaging way, and to ensure that they have a firm foundation for the narrative and game mechanics they are working with. Despite generally being excited for the activity, students sometimes struggle with getting farther than Student’s Hometown + Magic = New Thing…? This is where NotebookLM comes in.

Worldbuilding, the process of designing new cultures and environments for creative ventures, has been talked about in writing and education for decades now. From Tolkein’s work on “sub-creation” (Tolkein, 2008), to applying the architectural concept of design patterns to the creative process (Brierly, 2015), to using worldbuilding as a teaching aid and assessment tool (Hergenrader, 2014; McKenzie, 2023; Matuk, Herwich, & Amato, 2019) , to using worldbuilding as a media analysis tool, this process has been well-documented in its far-reaching ability to extend the creative process.

There are many approaches to worldbuilding, but for my game studies
courses I focus on the following categories:

  • Government systems
  • Economic systems
  • Magic systems
  • Culture/social structures
  • Environments

The students are invited to think outside of this list, and bring in aspects that help them make their world unique. By engaging in worldbuilding, students are forced to consider the interaction between complex systems, and they gain a deeper appreciation for the content we are covering. Using an AI tool helps to provide a creative space where the systems they are working with are contained, and they have the opportunity to explore and make connections in productive ways.

The Assignment

For this activity, NotebookLM supports instruction and student projects over the first half of the semester. Students engage with assigned material using one notebook that acts as the shared classroom “hub,” and they make their own notebooks as they research and develop the worlds for their games This includes all of the readings on worldbuilding for the semester. In the image below, I already generated a study guide, FAQ, discussion questions, and an audio overview (the podcast). Students that have access to the notebook can access all of the materials and use the AI tools to engage with course content.

Fig. 4 Class hub with generated content and sources circled

During the first week of the semester, students interact with the generated content in the class hub. The class hub ensures they are familiar with how the tool works, and it provides a space to introduce them to the ideas they need to complete the assignment. In the first week, they also start their own notebooks that they share with me and their research group (generally groups of three or four). Sharing a notebook, like sharing a Google Doc, is quite easy. We discuss how to use the notebooks as a centralized workspace for their research. We also set expectations for the process, which involves adding new material to the notebook each week (a website, article, or resource). This discussion creates a good opportunity to consider what kinds of resources are available for conducting research, how to select useful sources, and how to think ahead about the work they will produce. Within the first two weeks of the semester, students use the generative features to practice reviewing content. They read through the briefing document and annotate the study guide. They also listen to the audio overview for homework.

In class, I pass out physical copies of the article, and we do a jigsaw reading to discuss how that one fits into the overview. I split the class into four or five groups (depending on class size), and assign each group one section of the article. For example, Group One will take the introduction, Group Two will take the methods section, and so on. They are given roughly 15 minutes to read, and another 15 to annotate their section. Before beginning, I let them know that this will NOT be enough time to read the entire section, and to get through as much as they can (more on this later). Generally, their annotations will include highlighting main points and transitions, underlining or circling key/unfamiliar vocabulary, and margin annotations regarding rhetorical strategies, questions, and ideas for their work.

Fig. 5 Jigsaw Groups

When the annotation period is up, I ask them to discuss at their tables how far in the section they got, and how much of the overview they engaged with. These two things do not always correlate, but it gets them thinking about their reading practices, which is one goal of this exercise. This lasts about five minutes. The groups are then shuffled so that there is one person from each group at a table. Depending on the size of your class, you might end up with more groups than you had originally. The students discuss how the article fit into the generated overview and materials. Some of the questions I use to guide the conversation include:

  • What were the things left out of the overview that you think are important?
  • Did you find the study guide, podcast, or briefing document most helpful? Why?
  • Did you find any hallucinations? Which components of the article do you think led to those?
  • How would you podcast about this same topic?
Fig.6 Shuffled Groups

During week seven, the I use the class hub notebook to generate “midterm” multiple choice questions about worldbuilding, and the students generate analytical questions based on the articles in their folders. Using these questions, their task is to backward-engineer either a games course or a worldbuilding course that seeks to answer at least five of the questions. Aside from connecting their research to their genre and game subject, this part of the course also gives the students the opportunity to think critically about their use of AI in terms of the prompts and how well (and sometimes how poorly) the LLM structures the questions.

The students annotate their course with citations of what they used the AI for, as well as what changes they made to add value to the generated content. The peer review for this part of the assignment requires them to familiarize themselves with a partner’s research, and offer critical feedback on how well the research is reflected in their course.

The final step of this process comes in the weeks leading up to finals week. Their final project is a website that acts as a landing page, information hub, and teaser for their game idea. The students are asked to find a way to incorporate their individual notebooks into their website as an annotated bibliography that specifies how they took their ideas on their fictional world and extended them using research. They add a note as an introduction that reflects on all of the ways their research shows up in their final project, and their opinion on using the AI in class.

Throughout the course, students are asked to use NotebookLM to complete tasks like critical source analysis, adding sources for their classmates, brainstorming, note-taking, and reflections on close reading, creating detailed documentation of their research and creative process. All of this work leads up to the final deliverable of the worldbuilding website for their game, and by the end of the term their annotated bibliography is mostly complete. Also, utilizing NotebookLM throughout the course helps them pinpoint specific aspects of worldbuilding that they want to highlight in their own work. Instead of the “hometown + magic” or “GTA, but make it Tampa” models, I saw students branch out and try out new narrative genres and create unique game ideas. The use of NotebookLM helped students see into a wider scope of game development, as they were able and encouraged to engage with each other’s research as well as their own. This helped students who were not initially gamers get acquainted to other game genres and the ways narrative works in them, as well as helping them brainstorm what their game worlds could be (more than just guns and car chases). In the initial testing of this in class, I observed a wider variety of game genres, narratives, and worldbuilding elements in the final projects.

Fandom Research

A separate, but related, use of NotebookLM was developed to support a Fandom Research assignment in my Fan Studies course. For this course, the students complete a semester-long research project in which they apply a critical framework such as culture-centered analysis, visual analysis, or queer theory to their fandom research. The project encourages them to take a deep dive into the official and unofficial media of an established fan community, and it involves several key stages. They complete a research proposal, a revised proposal plus annotated bibliography, a research presentation, and a final major fan work (fan vids, fan fiction, etc.). The assignment sequence was redesigned with NotebookLM as a central tool in response to student feedback that the former fan work assignments felt disjointed, and it was hard to connect research to what they were creating.

The sequence for using NotebookLM starts same way as Games and Culture, in that we start by working in the class hub notebook and familiarizing ourselves with the tool. For this assignment sequence, though, we collaboratively create “satellite” notebooks that are based on fan categories: music, movies, TV, games, influencers, etc.

The satellite notebooks are not meant to function to support individual student projects, but are instead used to help students consider the unique qualities of various elements of an established fandom. We collect examples and research about how to analyze each type of media, and we use the notebooks to engage with and understand the theories we work with. Because students share access to the Notebooks, they are able to explore the same theories and concepts, but the generative tools give them space to ask their own questions and consider the topics in ways they find most relevant. As with Games and Culture, we jigsaw some of the articles in the class hub, and discuss NotebookLM’s generated content pertaining to it.

We regularly use the notebooks for in-class activities throughout the semester. For example, we do close reading with the articles in the notebook, generate research questions and practice prompt construction, and use the study guide to apply the course concepts to fan work they engage with during the course. This has helped to create a more solid through-line of how we analyze fandom with these frameworks, and how we can use the frameworks in the creation process. We also continue to generate and use the audio overviews as we analyze the construction of NotebookLM’s “podcast” compared with the construction of research-based fandom podcasts created by people. This semester (spring 2025) is the first use test in this course, so I am looking forward to finding out the outcome of using NotebookLM in their proposal’s annotated bibliography, on top of the in-class work.

Considerations

As I continue observations on NotebookLM’s efficacy in class, and have conversations with colleagues about AI usage in the classroom, I have come across some preliminary pros, cons, and considerations in thinking about this as a teaching tool.

On the positive side, this course organization allows students more opportunities to practice critical thinking and prompt construction. The cycle of generation → close reading → follow up → use allows you to keep the conversation of strategic use and adding value as the humans in the tech landscape at the forefront of the course. It helps the students stay in a collaborative mindset. I have seen better peer reviews as well, since they are keeping up with each other’s’ research through the linked notebooks. Lastly, it gives the instructor an inside, detailed look at how students engage with AI in their research process, a very valuable resource for monitoring and assessing progress.

This structure is not all rainbows and terabytes, however. Effectively implementing these activities requires planning and troubleshooting. Initial planning and setup is a time sink, at least the first time. Careful instructions and support time are needed to get everyone on the same page with the tech, uploading sources and sharing the notebooks, demonstrating how students can engage with the already generated content, and starting the close reading take a time and attention. I teach four credit classes, two hours twice a week. For me, two weeks gives me enough time to get the students acquainted with NotebookLM and practice the usage methods we will employ in the course. This will vary from instructor to instructor, of course, so if you adopt these strategies, be prepared to spend more time on setup than you may initially have planned.

Another concern is that, not everyone has (or wants) Google products. There are other tools that could be used to accomplish a similar workflow, but they require much more work on the part of the instructor and lack that central location that Google offers. Both of these together lead to the final drawback, which is that all of the pieces have to be there. This is a setup that takes forward planning and strategy to keep it going. It also requires students to engage the way you have asked them to. The initial week or two should offer time to practice this, but it does have the potential to go awry if you don’t have the solid structure and explicit usage guidelines from the outset.

“But won’t they just cheat and generate everything?” Possibly! But having clear policies on how you want them to engage with the tool does help with that. It also helps to have them document their AI usage as screenshots in an appendix to their assignments, and/or include citations of AI (that include their prompts) in their bibliographies. Nothing is a silver bullet in terms of making sure they use the tool exactly how we ask, but these are good first steps. If you do not already have AI statements in your rubrics, think of adding some in. Two areas that I have seen multiple professors use are Depth and Relevance. This targets the bland, sometimes tonally or logically incorrect writing of AI, and is an area where you can talk about adding value to the tool.

Finally, we have the ethical considerations. Credit, environmental impact, improper usage…the list seems to grow as fast as new AI tools come up. This is another area where discussion with students is key. It is also an area where you can add student agency to the course. Beginning with discussions about the ethical drawbacks to generative AI, and offering alternatives for students with strong moral objections (such as linked Dropbox folders, discussion areas in your LMS, etc.) can add the sense of agency while also affirming the students’ perspectives on the issues with these tools.

Conclusion

This course setup went from a half-baked idea a week before fall 2024 term started, to a system of research that aids production and collaboration in my classes. There are still bumps to smooth out, but crafting a framework that allows the class to be AI-inclusive has already had benefits for my students. As the AI landscape continues to grow, I hope to find other tools and systems that help to create research networks. NotebookLM works as a beginning system that allows accessible entry into complex topics, and the opportunity for students to practice critical technology use. It is my hope that as faculty continue to experiment and teach with AI, that we continue to refine these systems and expand the pedagogical landscape of AI use.

References

Brierly, N. T. (2015). Worldbuilding Design Patterns in the Works of JRR Tolkien. Proceedings of the 3rd Mythgard Institute Mythmoot.

Hergenrader, T. (2019). Collaborative worldbuilding for writers and gamers. Bloomsbury academic.

Matuk, C., Hurwich, T., & Amato, A. (2019). How science fiction worldbuilding supports students’ scientific explanation. Proceedings of FabLearn 2019, 193–196. https://doi.org/10.1145/3311890.3311925

McKenzie, B. (2023). Dungeons and Dragons and digital writing: A case study of worldbuilding. Journal of University Teaching and Learning Practice, 20(2), Article 2. https://doi.org/10.53761/1.20.02.10

Tolkien, J. R. R. (2008). On Fairy-stories. In D. A. Anderson & V. Flieger (Eds.), Tolkien On Fairy-Stories (pp. 27–84). HarperCollins.