Class 9 AI | Unit 4: Introduction to Generative AI – Creating with Machines
Unit 5: Introduction to Generative AI
From Predicting Data to Creating Masterpieces
1. What is Generative AI?
Traditional AI (Discriminative AI) is like a judge—it looks at data and decides what it is (e.g., "This is a cat"). Generative AI is like an artist—it takes what it has learned and creates something entirely new (e.g., "Draw a cat wearing a space suit").
2. Discriminative vs. Generative AI
| Feature | Discriminative AI | Generative AI |
|---|---|---|
| Goal | Categorize or Predict | Create New Content |
| Input | Data points | Prompts (Instructions) |
| Example | Spam Filter, Face ID | ChatGPT, Midjourney |
3. How does it work? (LLMs)
Generative AI for text uses Large Language Models (LLMs). These models are trained on billions of pages of text so they can predict the "next most likely word" in a sentence. It’s like a very advanced version of the "Autofill" on your phone keyboard!
4. Ethics of Generative AI
- Hallucinations: Sometimes AI gives confident but completely wrong answers.
- Deepfakes: Creating realistic but fake images/videos of people.
- Copyright: Who owns the art created by an AI?
🎨 Live Activity: AutoDraw
Experience a simple form of Generative AI! Google's AutoDraw uses machine learning to turn your rough sketches into professional icons in real-time.
Start Creating with AutoDraw 🚀Task: Draw a "Computer" and see how the AI suggests beautiful generated versions of your sketch!