A question to start with
You show a cat picture to a computer. It says “cat.” How did he do it? There is not a small character hidden in the screen. A program processed the image and proposed a response.
AI means artificial intelligence. This name brings together several ways to make a machine perform tasks: recognize an image, organize information or produce a response. There is not a single AI that does everything the same way.
A recipe and examples
Imagine two workshops. In the first, you follow a written recipe: take a sheet, fold it, then unfold it. The steps are decided in advance. It is an image to understand what an algorithm is: a series of instructions.
In the second workshop, many examples are shown in the program. He uses calculations to find regularities. This method is called machine learning. It’s a part of AI, not the only one.
A machine can thus propose the category of a photo. Other models can produce text, an image or a sound: we then speak of generative AI. Recognizing a cat and inventing a cat story are two different tasks.
What about the robot?
The robot is a machine that can act in the world, for example with an arm or wheels. AI is a set of computer methods. An AI can work without a robot in a computer. And a robot can simply follow pre-planned instructions.