NotebookLM on Project 2025

At the risk of ruining your Thanksgiving holiday and throwing us further into a “psychological funk” as we approach the political uncertainties of 2025, I’d like to share a brief (and successful) experiment I conducted yesterday using the full PDF of Project 2025 and the artificial intelligence (AI) model NotebookLM.

As you may be aware, NotebookLM is an experimental AI project by engineers at Google, collaborating with author and futurist Steven B. Johnson. I facilitated a webinar for the MediaEd Club of the Media Education Lab in October 2024 about NotebookLM which is archived and also linked on my “Webinars with Wes” website and my “Learning AI with Wes Fryer” website.

My AI experiment involved the following steps:

1. I downloaded the official, full PDF of Project 2025.
2. I logged into NotebookLM with my personal Google account. (It’s free, no paid versions / tiers are currently available.)
3. I uploaded the PDF to a new GoogleLM project, using it as my only “source.”
4. I clicked “customize” by “Deep Dive conversation” and entered a short prompt, asking NotebookLM to analyze the provided document and focus on impacts to our society, economy, political culture, etc.
5. After a few minutes, the AI audio conversation was finished and available. (Listen to it via my GoogleLM share link or via Google Drive. The first link requires a Google login, the second one does not.)
6. I listened to the audio, marveling at both the quality of the AI summary and analysis, as well as the deep pit of despair which we each risk falling into as we consider the content and potential impacts of the Project 2025 document.

After listening to the 15 minute summary and analysis of the official Project 2025 document, I created a second NotebookLM project and created an “audio deep dive” using the 48 page PDF “The People’s Guide to Project 2025” by “Democracy Forward,” which has also created the “Democracy 2025” project as a direct response to Project 2025 from The Heritage Foundation. That audio recording is 23 minutes long, but I actually found it to be a bit repetitive and not as concise.

Overall from an AI capability and analysis standpoint, I continued to be BLOWN AWAY by how good and insightful GoogleLM can be not only with its summary of key ideas, but also its analysis of complex, multi-layered topics. If you have not personally uploaded 1 or more documents to a GoogleLM project and created an audio podcast about it, you definitely should.

I read yesterday that ElevenLabs has updated its Reader App to create an even more audio podcasting tool than GoogleLM. I have not tried it yet.

As background for this AI experiment, I listened yesterday to a podcast I shared on my “Wes’ News Filter” public Facebook page. The podcast episode was “End of American Imperium: The Unsolvable Military Recruiting Crisis w/ US Major General Dennis Laich.” I shared a few of my reactions and thoughts about the themes of this podcast episode on a post on my News Filter Facebook page. One big thought was, “I need to read Project 2025 myself and spend some time learning about what it prescribes.”

Now is NOT the time for despair. It is the time for action. One of the responses to our era of political polarization and conflict, which I started last year, is the “Heal Our Culture” project. I started by collecting related resources and sharing them on a Google Site. I will continue building this project in the weeks and months to come.

Add NotebookLM to your AI toolbox. Please reach out and let me know what you do and learn! BlueSky is quickly becoming my favorite social media platform for conversation and idea sharing!