Google Cloud Holds 'NotebookLM Plus Training Session'
AI-Based Personalized Note Tool... Integrating Chatbot with Document Analysis
Analyzes Not Only Document Files, But Also Audio, Video, and Web Pages
"'NotebookLM' is a customized AI assistant that helps users gain insights more quickly based on the information they provide. It utilizes AI technology to offer tailored insights grounded on the data supplied by the user."
On the morning of the 6th, Jo Hyemin, Google Workspace Customer Engineering Lead Korea, introduced NotebookLM at the 'Let’s Talk AI: NotebookLM Plus Training Session' held at the Seoul Gangnam-gu office. She emphasized that it is a personalized tool based on artificial intelligence (AI) designed to enhance work efficiency.
Cho Hyemin, Google Workspace Customer Engineering Korea Lead, is introducing NotebookLM at the 'Let’s Talk AI: NotebookLM Plus Training Session' held on the morning of the 6th at the Gangnam-gu office building in Seoul. Photo by Lee Myunghwan
NotebookLM is a customized AI document drafting assistant launched by Google Cloud in June last year. It is a service similar to note-taking applications frequently used for work productivity, but with enhanced document-based analysis functions and chatbot-style AI. Essentially, after AI learns from documents, slide files, or voice recordings uploaded by the user, it analyzes the details and provides new ideas. Google’s large language model (LLM), Gemini 2.0, is utilized for data learning.
NotebookLM also serves as a hub where various work-related documents and data can be gathered and viewed in one place. It can analyze not only documents uploaded to Google Drive (Google Docs and Slides) but also videos posted on YouTube and websites. Users can directly upload their own voice files or PDF documents as well. Paid members using NotebookLM Plus can use sharing features, allowing multiple users to view a single note simultaneously.
NotebookLM quickly analyzes the collected work data based on Gemini 2.0. Document files can be rapidly summarized, extracting key points. Audio recordings and YouTube videos are transcribed into text, then similarly summarized for main content. As a generative AI-based service, it also proposes key analytical insights based on the uploaded data. This enables users to receive insights that assist with their work.
A distinctive feature of NotebookLM compared to other note-taking tools is its AI service provided in a chat window format. Users can ask questions as if using an AI chatbot, and the AI quickly responds based on the data uploaded by the user. For example, if a logistics cost table document with various conditions such as weight or distance is uploaded to NotebookLM, and the user asks in the chat window, "How much will it cost to ship cargo from Seoul to Busan?" the AI finds and provides the cost matching the conditions.
The chat style can also be customized by the user. Users can set the length of the answers and specify the focus of the responses, even by writing detailed instructions. For researchers, it is possible to configure the AI to respond mainly with academic content, or to provide simpler answers first to improve work comprehension.
The accuracy of the analysis has also been improved. The underlying model, Gemini 2.0, itself reduces hallucination phenomena where false information is provided, and NotebookLM further enhances trustworthiness by specifying the sources of information. Each piece of information displayed by the AI in the chat window is annotated, and clicking on these annotations shows excerpts from the original document’s source.
The actual performance of NotebookLM’s summarization and analysis, as experienced during the event, was sufficient for immediate work use. It took only a few seconds to analyze and summarize documents exceeding 10 pages, and it answered questions about the main content accurately. It also recommended new information in the chat window based on the document content. Since it is based on a conversational AI service, it is possible to analyze data naturally through dialogue. It even summarized the content of a recording of a presentation lasting over an hour immediately.
NotebookLM is provided free of charge to individual users. However, the maximum number of notebooks that can be created is limited to 100, and the number of data attachments per note is restricted to 50. In December last year, Google Cloud also launched NotebookLM Plus for paid customers, which increases the limits on notebooks and data attachments, and offers sharing features for collaboration as well as analytics on note usage.
Since NotebookLM is mainly used as a work tool, data security has been carefully considered. When using NotebookLM Plus with Gemini for Workspace for Google Workspace, the data uploaded by users and the Q&A content are not used to train the AI model.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
