Generative AI

KEY TAKEAWAY

Generative AI creates new content — text, images, music, code — by learning patterns from vast amounts of existing content.

💡 EVERYDAY ANALOGY — BEFORE THE TECH

🎨

Generative AI is like a musician who has listened to a million songs

A musician who has heard a million songs can compose original music in any style — jazz, classical, pop. They don't copy existing songs; they've absorbed the patterns of what makes music work. Generative AI has "listened" to billions of words, images and code samples, and can now create new, original content in any style.

👁️ Visual Explanation

Types of generative AI: Text (ChatGPT), Images (DALL-E, Midjourney), Code (GitHub Copilot), Audio (Suno), Video (Sora)

🌍 In Plain English

Traditional AI predicts and classifies. Generative AI creates. It learned from billions of pieces of human-made content and can now generate new text, images, code, music and video that never existed before.

Real-World Example

When you ask ChatGPT to write an email, it's not copying from existing emails. It's generating a new email token by token, based on patterns it learned from billions of documents.

🎓 For Students

Use AI to generate a first draft of an essay, summarise a research paper, or explain a concept in different ways until it clicks.

💼 For Professionals

Marketing teams use AI to generate 20 ad copy variants in seconds, then pick the best one — a task that used to take a week.

💬 Example Prompt to Try

"Write a professional email declining a meeting invitation, keeping it friendly and under 50 words."

🔧 Technical DepthOptional — expand when ready

Generative AI models are trained to predict the next token (word/pixel) given what came before. Large Language Models (LLMs) predict the next word. Diffusion models for images learn to remove noise from images to generate new ones. Both learn from enormous datasets using self-supervised learning.

Key Terms

TokenThe smallest unit of text an AI processes — roughly a word or part of a word

LLMLarge Language Model — an AI trained on vast text to generate and understand language

PromptThe input you give an AI — your question or instruction

⚠️ Common Misconception

Generative AI isn't looking things up or copying. It's generating new content from learned patterns — which is also why it can "hallucinate" (make things up confidently).