I am a subscriber of the Deloitte Monday Briefing; a weekly email summary of the UK and global economy. It is always an insightful read, and I was interested to see what NotebookLM would make of six weeks of emails from Ian Stewart, Deloitte's Chief Economist in the UK.
NotebookLM is a Google AI product. It allows you to upload sources to an AI notebook. For example, you could upload some PDFs and text files, and ask an AI to summarise, analyse, or discuss the content in a ChatGPT style interface. In addition, there are handy content generation tools. For example, you can listen to your material summarised in an audio podcast or generate a video overview of your content.
Copy 6 of Ian Stewart's economic summary emails into text files.
Upload the text files to NotebookLM.
Press Generate video.
As the wheel spun and my video was being generated, I became excited. What style it would be? What direction would it take? What gems would it uncover from the material?
The results were interesting. The video summarises the material with a slide show and commentary titled "The UK's £40 Billion Question". The presenter has divided the content up into five sections and kicks things off by introducing the current 40 Billion pound fiscal gap. Section 1 explains what a 'fiscal gap' is and summarises the challenges faced by Rachel Reeves. Section 2 attempts to summarise the big picture and has even included graphs as a visual aid (generated by numerical content found in the emails.) Sections 3 and 4 discuss global debts and the threat of AI while Section 5 attempts to round everything off by focusing on the metaphor of "plucking the geese".
This video has a jarring mix of cheesy discussion twinned with a doom and gloom outlook - and the accents are all over the place. That being said - it is visual, it quotes reliable sources, and it has helped me understand the content a bit better. Not bad if you consider I was able to create it at the click of a button.
If you are curious, the AI generated video is below, and it is about eight minutes long.
I think this is an example of where it can be helpful. I have easily created a visual aid to assist my learning. With that in mind, this has stopped me sitting down and reading the content myself, which is an important habit to maintain. It is worth noting that I could have generated an audio summary, to listen to in the car or walking to work.
The tech has enabled me to learn in a way that is convenient and efficient but could weaken my reading habits.
The above example highlights how I can use AI to generate realistic audio video content about anything I want - and this is something OpenAI hope to introduce to the social networking space with their new TikTok clone, Sora.
Imagine TikTok, but instead of users creating and sharing content by filming and uploading, they do this by generating. Type a few words, tag your friends, and see what video the AI spits out - then post it. When media shared online is overly doctored by software, lighting, and filters, it poses the threat of disconnecting the consumer from reality - particularly young people. It is worrying to think about what might happen to our general outlook when the content we consume becomes 100% not real. Will apps like this contribute to a mental health epidemic?
Article sources: My notebookLM.
For more reads like this one, join our mailing list.
Make your data safe, secure, and accessible with CSV Getter.