AI Fundamentals
BeginnerStart here. Understand what AI really is before you start building.
What is Artificial Intelligence?
AI is not magic. It is a tool that can learn patterns from data.
You have probably heard the term "AI" everywhere lately. Companies claim their products use AI, news articles warn about AI taking jobs, and social media is full of synthetic content. But what actually is AI?
Understanding AI is important because it is already affecting your life. The recommendations you see on YouTube, the spam filter in your email, and the autocomplete on your phone all use AI. By understanding how it works, you can use it more effectively and spot when it might be making mistakes.
- •Netflix recommends films based on what you have watched
- •Your email automatically filters spam
- •Your phone predicts the next word you will type
- •Self-driving cars recognise pedestrians and traffic lights
AI is software that can learn from examples instead of being told exactly what to do.
Traditional software follows rules that a programmer writes. For example, a calculator adds numbers because someone wrote code saying "take two numbers and add them together".
AI works differently. Instead of writing rules, you show the AI lots of examples and it figures out the patterns itself. Show it thousands of photos of cats and dogs, and it learns to tell them apart without anyone explaining what a cat looks like.
Think of it like teaching a child. You do not explain the rules of grammar before they learn to speak. They hear thousands of sentences and gradually figure out the patterns.
AI systems typically work in two phases:
- Training: The AI is shown millions of examples and adjusts its internal settings to get better at recognising patterns. This is like studying for an exam.
- Inference: The trained AI is given new inputs it has never seen and makes predictions based on the patterns it learned. This is like taking the exam.
The quality of an AI depends on the quality and quantity of training data. An AI trained on only black cats might not recognise a ginger cat. This is why data is so important.
Modern AI systems are built using neural networks, which are inspired by how the human brain works (though they are much simpler).
A neural network is made up of layers of "neurons" that pass information to each other. During training, the connections between neurons are adjusted to improve the output.
The most powerful AI systems today use a type of neural network called a Transformer, which is particularly good at understanding sequences like text. ChatGPT, Claude, and other large language models are all based on the Transformer architecture.