Author: Hannah Wind Date: Saturday, 28 September 2024
Location: Germany
Source: Private Collection (Screenshot of the website).
SUMMARY
Google NotebookLM (Language Model) is an AI-powered tool designed to help users with research and information organization. It is like a digital notebook that leverages advanced language models (like ChatGPT) to process and understand complex information from various documents. However, is it a useful step into an even more modernized world? What are the dangers that come with it?
BACKGROUND
Currently, there is a new trend on TikTok, promoting a new Artificial Intelligence website: Google NotebookLM. According to TikTok users, this new tool is “about to change studying for literally the entire world.” This is also what I thought when I tried it this morning: Instead of taking five hours to read around 75 pages and take notes of them, I could listen to the same content in just 17 minutes – stunning! But what is Google NotebookLM?
Google NotebookLM
Per an interesting article by B. Wang on Google Blog, NotebookLM is “an experimental product designed to use the power and promise of language models paired with your existing content to gain critical insights, faster […] a virtual research assistant that can summarize facts, explain complex ideas, and brainstorm new connections — all based on the sources you select.”
Podcast Feature
What people find particularly striking is the new podcast feature. Essentially, you can feed the tool a PDF text, and it can convert it into an audio file that sounds like a real podcast. The discussions are super engaging, sounding like two people are talking to each other, and the tool can simplify complicated passages into easy examples, making them very comprehensive for laymen.
What are the Dangers of Tools like Notebook LM?
Although all these features are great, some dangers immediately come to mind. As a law student, the first thing that rang my alarm bells was the copyright issue: There is a risk of unintentional copyright infringement, as NotebookLM might generate outputs based on copyrighted material without proper attribution. For instance, if I upload a book chapter from an e-book that I downloaded through my university, there is a risk that the tool does not designate the source as such. This goes hand in hand with sensitive information issues in the sense that users often upload documents containing confidential or sensitive data, which could entail data breaches.
However, the chain of issues with artificial intelligence tools like NotebookLM encompasses more than just the legal sphere. Still, it extends into the personal reality of people’s lives: Seeing as NotebookLM is so far only an experiment, not all information is accurate. While the service discloses this on the website, some users might trust the tool’s outputs without verifying, leading to the spread of inaccuracies in research, academic work, or business decisions.
Moreover, from personal experience, there is a serious risk that that using such tools might undermine genuine learning and comprehension, as students might rely on AI-generated summaries instead of engaging with the content themselves. Of course, it is tempting to just listen to a short podcast instead of doing the readings yourself – I have written earlier that I have done that too. Nonetheless, exclusively using the tool comes with a significantly different learning experience that is not necessarily to a student’s advantage.
Overall, while I find the tool helpful support, it must be enjoyed with caution. Caution to one’s own experience of learning and accuracy of information intake, caution regarding other people’s and one's data caution in terms of copyrights.
REFERENCES
Biao Wang. (11.09.2024). NotebookLM now lets you listen to a conversation about your sources. https://blog.google/technology/ai/notebooklm-audio-overviews/
TikTok Video by leahmiami (27.09.2024). https://www.tiktok.com/@leahmiami/video/7419141961753300267?_r=1&_t=8q62MJgCnvM
Access URL here: https://af52232c-237c-4b2d-8b28-6c803d233b21.usrfiles.com/ugd/af5223_abc319cbab8f4f31b8babe267a2c94f1.pdf
Access PDF here:
Comments