1.3 Evolution of AI

AI has evolved through three broad stages. First, rule-based (symbolic) systems where humans wrote every rule by hand. Second, machine learning, where systems learn patterns from data instead of being explicitly programmed. Third, deep learning and generative AI, where layered neural networks learn rich patterns from massive datasets and can even create new text, images and code.

Think of teaching a child to recognise teff versus rice. A rule-based approach would list every difference by hand. A learning approach instead shows the child thousands of labelled photos until it figures out the patterns itself — which is far more flexible.