What are AI, Machine Learning, and Generative AI?
Think of it as a hierarchy of concepts, simplified for everyday understanding:
- AI (Artificial Intelligence): The broad field of creating computers that can do tasks typically requiring human intelligence (e.g., recognizing patterns, making decisions).
- Machine Learning (ML): A subset of AI where computers learn from data rather than being explicitly programmed for every scenario. It’s like teaching by example.
- Generative AI: The newest branch that doesn't just analyze data but creates new content—like text, images, or even code—based on what it has learned.
What Large Language Models (LLMs) can and cannot do?
✅ CAN DO
- Summarize long research papers.
- Draft emails or library descriptions.
- Translate metadata across languages.
- Brainstorm project ideas.
❌ CANNOT DO
- "Think" or feel (they predict the next word).
- Guarantee 100% accuracy (they can "hallucinate").
- Access real-time info without external tools.
- Possess human-level common sense.
Overview of LLM-based Agents and Tool-Augmented AI
Standard LLMs are like smart consultants; AI Agents are like pro-active team members. They don't just talk—they take action.
Tool-Augmented AI refers to systems that give LLMs "superpowers" by allowing them to use calculators, search the web, or access library databases directly to perform specific tasks autonomously.
GPT-style Systems vs. Notebook-based Assistants
GPT-style (e.g., ChatGPT): A conversational interface designed for general-purpose interaction across any topic. Great for drafting and brainstorming.
Notebook-based (e.g., NotebookLM): Designed specifically for grounding AI in your own data. You upload your specific library policies or research PDFs, and the AI only answers based on those specific documents, providing citations for its claims.
Multimodal AI: Text, Image, PDF, and Web
Modern AI is no longer limited to text. Multimodal AI can "see" and "hear." It can analyze a scanned PDF of a historical document, describe what's in a library photograph, or browse a website to summarize its content—all in one session.
How these tools appear in Library Workflows?
- Discovery: AI-powered search that understands intent rather than just keywords.
- Metadata: Automatically suggesting subject headings or cleaning OCR errors in digitized collections.
- Reference: Supporting chat services with instant summaries of library policies or FAQs.
- Administration: Analyzing monthly usage data trends or drafting outreach materials.
More Details
For more in-depth information and detailed concepts, please review the supplementary slide deck.
View Module 1 Slides →Video Tutorials
An introduction to mainstream LLM model interfaces, using Gemini as a primary example to explore standard chat features and settings.
A deep dive into Gemini's Memory and Retrieval capabilities, showing how the model manages long-term context and retrieves specific information from past interactions.
Explore what NotebookLM is and how its unique "grounded" interface differs from standard chatbots by focusing on your specific documents.
Learn about NotebookLM's Studio features, including how to automatically generate mindmaps, interactive podcasts, and structured study guides from your source library.