Generative AI Basics
What Makes AI 'Generative'?
Classical AI classifies: is this email spam, is this tumour malignant? Generative AI creates: write this email, draw this picture, compose this melody. The shift from analysing to creating is what makes this generation of AI feel so different.
Key Concepts
- Discriminative models assign labels; generative models produce new samples
- Text, images, audio, video and code can all be generated today
- Generative outputs are new combinations of learned patterns — not copies from a database
- Quality has crossed a threshold: much generated content passes as human-made
- The same capability that drafts your report can also fabricate misinformation — capability and risk are twins
Try It Yourself
Write down three tasks from your week and label each as classification (analysing) or generation (creating). Which could generative AI have accelerated?
Lessons
▶️ 1. What Makes AI 'Generative'?
▶️ 2. Inside a Large Language Model
▶️ 3. Working with Chat AI: A Practical…
▶️ 4. Image, Audio and Code Generation
▶️ 5. Hallucinations, Bias and Verification
▶️ 6. Generative AI at Work and School
▶️ 7. Choosing and Combining Generative Tools
▶️ 8. The Near Future: Agents, Multimodality…
Course Home